/agdg/ + /vm/ ~ Amateur Gamedev General

Anton and Coolpecker is programmed fairly well. Look up the source code for that perhaps.

This is the "STOP FUCKING PROCRASTINATING AND GO MAKE YOUR FUCKING GAME" EDITION
Start up your favorite IDE/Engine and get to work now.

Does that mean the topology is good enough?

On a side note, does anyone feel like making levels in Blender or drag-n-drop via Unity, kinda clunky?

I used to use Source Engine back in the day. Say what you will about BSP's, but they're pretty tight, and allow for easy distinction between two different areas (say the interior of a house being in the same level as the outdoors for instance) and overall really modual geometry.

How do you guys approach the execution of level design?

I redlined your traffi jam.

This looks better.

I would add a cut in around the shoulder / armpit. Or if low poly is key then Slide / tilt a ring from the arm back towards the pit:
____
| | | | | | | | -> digits are that way.
|

to
_____
|/ / | | ||| | | …
|

Wherever you have a bend you want more rings / quads, like how the quads pinch together in the groin, that's good.

You may want to arrange for them to get close together again at the elbow too. Imagine the elbow bending and straightening. When it's bent you don't want the quads there to get too big and chucky looking.

Try viewing without the smooth modifier – raw flat polygons. That will reveal sharp edges. Any too sharp / high contrast angles you'll want to adjust the lines closer together to resolve.

how about instead i work on the thing i actually want to make, and make that instead

I don't quite get it, so I move some rings to the elbows and slide a few vertices to the armpit?

Now bouncing between 8-19% CPU usage on linux, thanks user at

Nice running animation, I suck at those.

I'm probably going to break your heart, then.

That was my first running animation ever. This whole project is my first attempt at animations and 8-bit pixel art.

Like this?

As long as the environmental sprites are better.
I can't stress this enough:
DO NOT HAVE BLOCKY ENVIRONMENTS IN A PIXEL ART GAME
DO NOT HAVE BLOCKY ENVIRONMENTS IN A PIXEL ART GAME
DO NOT HAVE BLOCKY ENVIRONMENTS IN A PIXEL ART GAME
DO NOT HAVE BLOCKY ENVIRONMENTS IN A PIXEL ART GAME
Otherwise you end up with uninspired looking pixelshit.

Examples?

That'll do, pig.

no prob. p.s. I had a bitch of a time with GLX too, it's not just you. Also, I love valgrind memcheck for finding weird segfaults and uninitialized memory.

Finally! thanks, I only have to make the foot now.

No hard feelings m8, keep up the good work

Of what? Blocky environments or good pixel art?

How do I into art/character design? I have a great idea for a game and I have the capability to build it since I've studied computer science for 3 years so my coding is good, but I'm no artist and the idea rides on good character designs.

Blocky environments.

Focus on building the gameplay first, then we can worry about the art.

what did he mean by this

Thought that it might be a common experience. Its now actually running almost correctly in Linux (some minor housekeeping needed) so that's pretty nice. It seems to be a pretty fast port, although I've never ported anything before.

Depends on what kind of characters and arts.

Just use shitty art / placeholder graphics for now. Prototype your game while you're trying to get gud.

You can get some 3D assets, load 'em up in a modeling program, then modify them or use them as your art subjects, where you can control all the lighting, composition and etc.

Porn is also cheap to use as references. Practice makes perfect.

I have ideas, but haven't fully decided on an engine. The ones you pay out of pocket or give royalty dues are probably easiest to use, but that's a pain, so I'm looking at open source since it seems more worth it. What are good open source engines for commercial indies? Is Godot my best bet?

I'm not having any issues actually programming this, I just don't understand how I can simulate water pressure on a grid. My program simulates that in a very hacky way (essentially finding and remembering a path to a source block or a block with the highest concentration of water, and drawing from it) but I need to actually calculate pressure.

The game I want to make is a 2D JRPG so no need to learn modelling or anything. I just have like 0 original ability, but I'm fairly good at copying things that others have made.

For 2D or 3D vidya? Also, general purpose engine or specialized?

Begin using simple sprites to refine mechanics.

Just copy ff5 characters and edit them later.

Wait until Godot 3 comes out for 3D. 2D is pretty good though.

Pressure = density / area

You have units of space. In order to have pressure you need compressible fluids, i.e., you need to be able to have variable amount of unit water per cell. If the cells are 1 unit water per cell, they have the same pressure in every cell.

If you have more fluid per unit space, you don't have to compute the pressure by counting cells above a cell. You can just count the amount of fluid in a single cell. In fact, you can have a value called: "waters" in each cell that count how much waters are there. More waters more pressure.

Pressure is just how much matter there is in an area. Movement left and right is just water trying to equalize counts between adjacent cells, flowing from higher to lower counts. Treat each cell as a bucket for water units that can be 0,1,2,3,4, etc.

When a water moves into a cell add the direction it's moving as a velocity vector into a cell's vector per cell. Divide the velocity by friction each pass, and add the constant gravity of 1g to the result. Move N waters per pass by the cell's velocity vector. If blocked from movement in a direction, negate the vector and move the other way (unless also blocked).

If you want to compute pressure just total the absolute value of the vectors in a given region.

The better question is: why do you need to compute pressure? Add a sensor to the simulation and record the forces it encounters.

A drop / cell of water doesn't know anything about the object it is in except for its own properties and immediate surroundings.

The first idea I want to start with is 2D, the rest is probably 2D with possibility of simple 3D if I feel comfortable enough. The specialized engines I've looked at don't seem to have what I'm looking for (I mainly want gameplay similar to classic Zelda or Mana).

The better question is: why do you need to compute pressure? Add a sensor to the simulation and record the forces it encounters.

Oh, I don't need the pressure of the body of water, just the pressure of each tile. I plan on adding shit like pumps that affect the flow of water by moving pressure from one side to another, so my earlier hack was useless.

I'm assuming I should also keep a buffer of the water tiles for the calculations, so each tile operates on the 'old' data and then switches it out at the very end, right?

Thanks

I'm not saying use 3D models in your game. I'm saying use them to make the poses you can "copy" as 2D sprites… or, you can even render the 3D models to 2D and use them as sprites.

Ultima Online used 3D models for much of it's 2D RPG graphics, then later came out with a 3D client, but most people preferred the 2D assets as they were rendered with a high res model and detail textures, then touched up by hand.

It's just a suggestion as a crutch for coderfags who can't into art just yet. Generating 2D from 3D was the dirty little secret used extensively in the 90's / 16bit console era to generate graphics like crazy and it's still secretly used today. You can tweak, edit, recolor a character that's a 3D model a lot faster than you can someone else's 2D sprites… Some of the beloved "pixel art" began as shitty flat shaded 3D crap, and was then rendered to 2D placeholder sprites while in production. Once the number of frames and poses were nailed down, the base rendered output was retouched by artists to make consistent 2D sprites. Not every game was like this, but many used the technique.

Start with stick people if you must.

Wouldn't something like Solarus be perfect for that? solarus-games.org/

Sounds good.

A water pump with such a cellular automata sim just moves units of water from the local area to the next area. The simulation does the rest by filling the vacuum. To make a pump just select cells and add a constant acceleration vector to those cells when you add the gravity. The simulation will do the rest – including back up and flow the wrong way if the pressure on the other side is great enough.

make your game ENTIRELY out of free assets made by others

Worked for me, I didn't actually make any art for my last game, just programmed it all

Not sure if it's flexible enough for me since all the examples the site lists are ALttP-clone fan-games. Are there notable works made in this engine that aren't strictly based on this style? (Godot's simple multi-platform deployment also seems like it can potentially be put to good use down the line, but that thinks more ahead than I probably should…)

None that I know of, and I haven't looked at the docs much so I don't know exactly how flexible Solarus is.
Godot is nice for 2D vidya if you like its hierarchal scene tree and IDE. If you don't like the architectural or want even more freedom, there's lots of options. Orx ( orx-project.org/ ) is neat if you want to try data-driven game development, MonoGame is basically XNA 5 if you're into that, LÖVE is neat for luafags and Pygame for Pythonfags. There's also the Anura engine used to develop Frogatto ( github.com/anura-engine/anura , docs at github.com/frogatto/frogatto/wiki) and software libraries like Allegro if you want something lower-level without going to SDL or writing your game against platform-specific APIs.
There's also a bunch of Haxe frameworks and engines out there but you'd probably want to ask the Speebot dev about them whenever he's on.

Hello devs!

I'm the drawfag who proposed an /agdg/-tan design a few threads back. Making use of previous suggestions, I chugged on and came up with this.

Some notes:
- Changed the sweater to a blouse because it resembled Danielle Deepfreeze.
- Added pantyhose because previous design resembles Vita-tan.
- Added the color scheme of the /agdg/ logo and other symbols, but kept it nice and clean so it wouldn't stand out too much.
- Still an ara, and tried making her a bit THICC-er than the previous one. Refined the linework too.

From one creator to another, good luck anons.

I think you just gave me autism

>>>Holla Forums
>>>/oven/
>>>/bog/
pick your poison

Just how viable are multi mesh models like 5th gen 3D console games on modern engines? Is Unity or Unreal going to freak out because the mesh isn't being deformed?

JESUS CHRIST THIS LOOKS LIKE BRIANNA WU TIER GARBAGE PUT IT IN THE FUCKING GAS CHAMBER

reminds me of this thing

Don't mind the haters, looks fine to me, and I hate western art with a passion.

Nah man
Just nah

Dropped in just to say that this looks like shit

At least it doesn't have a red nose

kek

the engines are probably not going to freak out because the mesh isnt deforming.
they are probably going to freak out because of the shit you will pull so you dont have to use normal animations.
why dont you weightpaint it in a way that the single mesh will behave like its made up of several?

I expected as much. This isn't a hugbox like DA or Tumblr.

But hey, -tans live and die by the people who adopt them. I put it up for people and if they want it, it's all theirs. If not, I still had time well-spent on Character Design. (Different ID, working on a different computer, btw)

Make it a lazy loli instead

But isn't that ironic? The whole point of /agdg/ is getting stuff done and making your own vidya.

…Or is the irony the whole point?

I wouldn't say lazy.
I'd picture it as a smug, yet motivational loli, that's overworked and tired.
You know how unkempt Vivian is? Take that to the extreme.
She'd make fun of anons for using Game Maker, and such I use GMS, and generally be smug about it. And yet, despite being frustrated and tired, and no matter what they might use, she'll always be there to motivate all the anons here to just, like, make their fucking games.

It should be fine, because that's how robots work, but joints will probably be noticeable unless your render style is similar to cell shading, or your clothing design masks it very well.

One thing to beware of is that a tree of matrices will typically require re-normalization over time. That means every so often you regenerate the matrices from the radians/angles.

Quaternions are a good fit for rigid-body systems because they can be rotated and interpolated easily and renormalized to a unit length of 1 (one) very quickly; They should be converted into a matrix before transforming mesh vertices's with them though (since it's slower to transform a point with quaternion than matrix).

Remember, on a humanoid the "root node" of the limb hierarchy is the crotch. Aussies know what's up. The camera usually had its own a fixed offset directly above root rather than derive camera position to properly be between the eyes. The rounding errors accumulate through the matrix tree. That's not noticeable on entity animations but camera would get jittery if it was more than a few nodes deep in the tree.

That said, unless you've got a ready-made rigid body system at your disposal already, then it'll probably be less work to just weight your vertexes. You can even make a multi-part robot with a single deforming mesh if you model it right.

Fucking whoever put that sign up doesn't want me in their cave. I bet it's full of treasure

Sure is a lot of anti board-tan shilling all of a sudden, especially considering all the constructive discussion that was had at the beginning.

I like it but you misspelled /agdg/-tan as /adgd/-tan.

When i made the general with /agdg/-tan as the edition, a bunch of people were throwing a bitch fit.
Personally, I never changed my opinion. I said the same thing here as I did when the idea first came about.
I'm just not fond of the design. It's not terrible in general, but it's not fitting of my idea of /agdg/.

i often accidentally spell it /adgd/ too

Alright, I'll take that into consideration for another design. I had two competing ideas because there were people who wanted an ara and another that wanted a loli.

No wonder someone posted Watamote as reference.

Fuck I even checked that title three times and I still got it wrong!

Let me dig up that other idea I was working on. I might transfer some visual elements to the new design, but yeah, I got the gist of what you're looking for. I'm having way too much fun to drop it now.

Keep it going, man. Obviously, it's impossible to please everyone, but whatever you make will be better than Doggo, at least. Fuck that thing.

I just finished her.
critiques please

Tiny feet and flat colors

l-lewd!!

Aren't those shoulders a bit broad?

You don't come here often do you?

What about a lazi loli that often turns the motivation switch on and devs for months, then drops the project, laze arround, gets motivated again, work, drops, into an endless cycle of nodev
No, more like Anzu or Sachiko, it has to be cute

Remember. Lolis can into oppai.

Even if flat as JUSTICE, a motivational loli would sometimes stuff an oversized bra to cheer for and motivate her nodev.

The skirt looks bad, also thighs can be thicker
The face looks weird, i think the cranium is too big, or the face is too small

...

Not if he's a guy .

She allows you to headpat her

love the art style, and updated design.
face is still wu'ish though.
Really w/e design you end up with I'd probably like it.
You should definitely feel incentivized to take creative control.

As long as it's not some shitty moe character that's mostly fine.

yes, but that rarely happens; generally it's incremental stuff (a piece apart of a grander whole).

More like, getting something partway done, then switching to something else, and rinse repeat for eternity.

Don't use flat shading (only ambient light); it makes anything look 1000% worse than it actually is.
Utilize at least semi-realistic/anime-inspired shading. picture related.
Feet are too small, but otherwise proportions look almost correct for a loli. use marvelous designer for clothes that don't look like shit.

That is actually surprisingly accurate.
The only way I could see /agdg/-tan as a loli though is that we're exemplifying the unmature aspect that gamedev requires.
So, the essence of how one has to handle their time, and the point that… to get anything worthwhile done one has to abandon what other people consider "mature priorities" for gamedev.
Also the point that developing a game is more than likely a dream one has had since they were a kid though that could just be me, may just be projecting that part.

as long as it's not overkill, but I always liked hestia's design so that would be legit.

an /agdg/-tan loli would be for headpats only

y-yeah, t-that's what i meant…

It has been mine since i was 8, im 29 now and i finally said fuck this and i kept myself motivated for over 7 months already, im confident i can finishmy dream game this time.

Hey guys

How about we

Guys

How about

Guys listen

What if we just

Posted progress instead of trying to make an autistic OC?

...

im slacking at my crappy dayjob right now, can't
gonna be diopping it in 9 days to go full indie

This, its like that user that actually bought an Ouya because of ouya-tan even though the pictures had "please don't actually buy an ouya" on them, Tans are have real power

dropping*
fuck, need more coffee

once you name something you give it power.

may raise productivity actually.
Also, the whole goal of /agdg/ is producing "autistic OC".

I have a similar thing, though I'm 27, and working on dev for… fuck almost 4 years now; current project for about ~2yrs of that time.

I believe in you

Unless you are going on autismbux to fund your dev time I got bad news for you

No amount of a forced loli waifu would make people be more motivated. Pity and motivation are not comparable. Now post progress or fuck off.

Happy to oblige. You seem to have this belief that I can't work on conversation mapping and also occasionally discuss something kinda cool.

t-thanks

The initial idea was doing kickstarter, but i think the Star Citizen crap was the last straw, simply mentioning it is a bad idea, instead i will open a patreon after i release a demo (3~4 months from now), i have savings to stay afloat.for a year until the patreon raises to a good level about 800 bucks/month would be great

you are a baby
I'm scripting, wanking, watching a tv series and discussing with you people at the same time

How many hands do you have?

...

fixed somethings

I'll have to look into that more.
I'm trying to emulate those pokemon 3ds models. I never touched an engine so I don't know what the lighting will be like.

Put the model on the engine before doing final touchs unless you are fixing the mesh
Even a super compatible importing like Blender -> Unity changes the appearance of the materials a lot

If there was an agdg-tan to model I would probably make some progress using agdg-tan as a base for temp assets.

Having such assets freely available maybe coderfags would have less excuse not to make game while trying to art… but I might be wrong about that.

The progress I'm making isn't really post-worthy since right now its just a bunch of really fucking ugly C macros and structs that normalize pthreads and Win32API threads without needing a wrapper function. This is accomplished via some crazy shit with generating typedefs and structs via macros, and using only the lowest common denominator features of both APIs. Most of the time is spent on a whiteboard thinking up the strategy, much less time implementing code. When I'm done I'll have the leanest meanest cross platform abstraction layer to build the engine on.

If watersim dev doesn't hurry up and get some water pumping going on I might make a watersim in textmode as an example and to smoke test the threading with batch processing, and to show some visual progress. Wrote a paper on water sim once upon a time, but it used particles not voxels – and my goal was accuracy for a commercial sim, not real-time gamedev.

Your legs are a bit weirdshaped.

looks like cellshading effect, and diffuse + ambient lighting.

thats a nice idea, i would put her on my game if she has a great design like Vivian

Are you using the face as a plate, or is it part of a head mesh? Looks like you need to recompute your surface normals, may also be why legs have weird line, or that could be your texture though.

Some modeling tools get weird about surfaces like face plates. If a plate, then extrude the edge of the plate backwards a good amount, and then recompute normals. Then erase the extruded part, leaving the normals along the face plate's edge with proper normals.

I actually rather like her. For those who don't like this design, I'm sure it would evolve over time… let's not forget that the original Vivian design was pretty tumblry.

Apologies for more autistic OC.

Are you guys who have so much time to devote to making drawings of thread characterizations actually working on any games? Maybe you could post about them instead.

okay that's kinda cute.

Making games is hard, pushing problem hair tumblr OCs is not.

I've been wondering, user. Does the program you use export in any format you can actually use in a game, or are you going to have to go through everything a second time for that?

I'm doing story shit, user, and will spend the next several months, if not more, on it.
I post snipits, but I can't post much.

That, and working on story alone is fucking hell, and it's nice to take a break and talk about something nice like this.

I'll have to rewrite it all by hand.
It exports as an image.
It's dumb, but it'll let me recheck for spelling mistakes and shit.
This is more a visual guide to make everything easier for myself.

Try again.

...

After careful consideration, here's my critique:

Do you wanna work for vidya or starbucks?

Consider suicide degenerate.

Inflation and fat are two very different things
Both very shit
But very different

Why don't you post your torso if it's so thin, then?

I think the main problem is the hair
A simple blonde/black ponytail or long straight hair (with tiara) would be thousand times better

BUT WE HAVE TO DESTROY THE CAPITALIST PATRIARCHY YOU GUYS

that's not how thighs work
that's wrong

...

R8 my MP5. Tried to keep it under 500 as usual, but sadly the MP5 is a very cylinder-heavy gun, and cylinders eat polys like /cuteboys/ eat cock.

Why even live.

That definitely looks like a gun, good job.

I recommend HaxeFlixel for 2d vidya.
It's a FOSS game framework that targets every platform, is written and maintained by competent people, gives you a lot of power while being useful, and is pretty mature by now. A lot of commercial games were made with it. haxeflixel.com/showcase/
If you want to go lower, OpenFL is the next best thing (HaxeFlixel is based on OpenFL). Even lower - Lime (OpenFL is based on Lime). My engine YUME uses Lime.

...

...

...

This week I've finished world 3 of Speebot. That's 150 levels done out of 200. Best case scenario - I'll release in June/July. Time to shill my greenlight again: steamcommunity.com/sharedfiles/filedetails/?id=924909034

ahh shit, user, congrats.
have a headpat for being a good boy
I for some reason was expecting 8 worlds or something, never thought there'd be so many levels per world.

It might just be because I work so slowly, but you're certainly cranking those levels out. You sure you're putting enough thought into the designs of the levels? I don't doubt you per se, just looking out for ya.

Does this anger you, user?

Reminds me a lot of scratch
scratch.mit.edu/projects/19708756/

Thanks user. I'm taking level design seriously, it's the central point of the game. I hope you'll enjoy it!

Someone paid $100 to shitpost, let them. I'm more annoyed by Not Minecraft above it.

Ughhghhhhhghghghghgh

Can someone point me to a good tutorial for modeling in blender? I was following this:
cloud.blender.org/p/game-asset-creation/56041550044a2a00d0d7e073
But

Get back to work, or I'll call you a nodev!

just like make game

This is why we need a cute tan, to motivate lazy fucks like that user

So I've spent a coupla months now working on a 2.5D platformer for my college thesis it's in animation, not in """""game design"""" at least and this is what I have to show for it. Now that I'm finally done, and I'm not being held to a strict deadline to have this thing finished, I'm thinking maybe I should just scrap 90% of what I have and start from scratch.
Some of the models and animations I think I can still use, though I wanna improve the textures in the end. And the code definitely needs to be overhauled entirely; you get stuck on walls and corners, and you can see the framerate take heavy hits the moment you step into a building because the damn thing is made out of hundreds of textured squares being rendered at once.

otherwise, how does it look

Lost many a project that way before version control.

welp.
C:\SPACEWAR>del *.BAS

At the very least, zip archive before every big change.

The kick is way too clunky, change it into a slide/low-kick that lets you keep momentum, never force your player to stop in a platformer, unless you want to how him a mechanic or something.
also, instead of a health bar make your sheep character lose wool every time it takes damage, or even better change into a cute sheepgirl that gets embarassed when she is half naked
Also using the sheep/wool theme you would steal a few ideas from the kirby yarn and yoshi woolly games

But all the artfags are too busy making game

What grade did you get? I'd have given you a D- for the protag, but the scuttling enemies bring it up to a solid C. ofc knowing "academia" these days that would get curved to A

How much do you plan on charging for this when it's released, user? Because provided it does not have Denuvo, that shit is fucking cancer I will most likely buy it, and then if it delivers, buy copies for my friends almost immediately on launch.

I feel like I want to kill myself
I'm also not gonna be able to make something that was this clean by myself

It's gonna look shit, it's gonna work like shit and it's gonna make me waste a ton of time

Damn user, don't lose hope
At least you got experience out of it

Thanks user
I mean I can't really lose hope at this point, the guy I was working with has been making animations for a week, it'd be pretty gay of me to drop out

I just need to get some rebound, fuck if I know
This does feel like super shit though

what prevents you from making a trigger that follows the current player position?

Around 15-20 bucks, with a 25% launch discount (or something like that). And it's DRM-free of course. Will likely release on Humble first, Steam - second (if I get greenlit). Does that sound reasonable?

Tell him you need a week off, then go read success stories about devs who did it like the Stardew Valley guy, if a guy like that can make a game so can you.

How would this work?

Try to get it on GOG, too.
gog.com/indie

Humble first?
Weird choice

This is just so frustrating to look at for some reason. How many years of college?

You're gonna laugh: I didn't even get around to coding momentum changes when kicking. You can keep on sliding across the floor while throwing kicks. I've been trying to scrapbook what kind of techniques she'd have, flip kicks, dive kicks, debating between low jump kicks or slide kicks when running, shit like that. That little sheep following her around was also supposed to be a Nier-style alt fire, but I was out of time by then.
Aside from all that, I'm also trying to decide just what I wanna do for graphics. Stick with the 3D models, switch to HD sprites, why the fuck does Maya keep crashing the moment I open it all sorts of decisions. Only thing I'm certain of is I need to kill the Unity stock lighting.
There already is a plan for a half-naked embarrassed sheepgirl mechanic, but not for showing health.

I would like to know what you'd adjust about her design, since I've been struggling to make it more interesting myself. I'd show you a proper T-pose, but Maya crashes.

One semester planning and modeling, one semester animating and coding.

Is it a robot/android? Make her a biological creature and make her fluffy keep human face, no yiffs
Also download Blender

Will do, hope they accept it.

It's the friendliest platform I've seen yet, they give you a widget to sell your game on your own site and take a really small cut (5%) compared to other distributors. They give you the option to host a free demo and everything, too.

She's a cybernetic thing. The whole thing is loosely based on Do Androids Dream of Electric Sheep.

Can it actually be a thesis if you've only been in college for a year?

Nah, that's just how long I spent working on this project. It was a two year bachelor's.

This is a waste

So ditch the mid-poly look and make her fluff bigger and rounder, is what you're suggesting?

...

do i really need to spell it to you?

is better if you figure it by yourself but in case you don't
instead of directly using the on trigger events between the character and an interactive object you have three objects. 1 your character, 2 a trigger that updates its position to 1's positon every frame and 3 the object that you need to interact with. then use the on trigger events from 2 to send signals to 1 and 3

Big wool sweater with long fluffy hair

Nigger I said that all the OnTrigger events are broken
What you're basically telling me to do is just add another item in the broken function for the fuck of it

yes

I still think you should link the demo on there. Also, don't you get a graph or something of how many votes you got? Can you share that legally, because I kinda remember steam being a bitch about that?

You do realize putting Denuvo in your game costs hundreds of thousands of dollars, right?

post the code then, odds are is just some dumb shit like most agdg problems

Very.

Well aware, but nowadays it seems like a good chunk of the notable shitty indie games end up using it because they can't handle muh criticism. Just covering my bases ahead of time.

It's like 10 different scripts
I doubt anybody here would be able to fix it

If you wanna have a shot at it it's the Super Character Controller

Do you have a snapshot before broken, and after?

Because, then it's just a diff to see what changed, pair down to isolate, and then engineer a workaround / fix.

Seriously, back shit up always, version control if possible.

Oh no you misunderstand, it wasn't me that broke it.
The script itself broke itself at some point before I downloaded it

This ain't me

Jesus fuck dude

What would be the best way to categorize and detect classes in UE4? Apart from casting/class checks (which might require me to make some empty classes just to have others inherit from them) and tags, are there any other ways?

Wouldn't even be a question if GameplayTags were standard in UE4 but as of now I'd have to modify UObject/AActor to use them as default instead of the usual string tags, and who knows when they will make them standard.

yeah, you made it pretty clear what you were saying from the very first post, the guy you're responding to is just goddamn brick wall

Still looking for a way to sort objects by position in C#. There's about 2,000 objects at max, and their position is a ushort, so it should be easy.

I notice there's a SortedDict and SortedList class, but I need to iterate them in reverse order, but they want the key, instead of the index. Otherwise, it does not make sense to build a stack of 2000 items 60 times a second just to do this.

The demo is out of date now, I'll need to make a new one soon.
I've got 370+ "Yes" votes right now and 28 comments, which all seem to be pretty positive. Speebot seems to be doing much better than my first game (Hypnorain), which got accepted anyway.
Unfortunately I don't know that much about the process, even though I see the stats. My guess is as good as anyone's, but I think Valve is looking for games that seem polished enough and are not universally hated. I've heard from other developers that "No" votes don't matter at all, because Valve recognizes that niche games target specific minority audiences.
If there's anything else you'd like to know, I'll try to answer if I can.

Use the correct data structures designed for spatial data (positions).
Those being spatial data structures. I'm partial to the octree, but if u use 2D there's always the quad tree.
If you're overly concerned about speed, then use a linearized version of the quad/octree with Z-ordering; thus having direct access to nodes via a dictionary w/o having to recursively iterate from the root node.

I just used a Sorted List, since I can get a list of values from it, which is index based the way I want it to. C# doesn't have any quad trees by default, and even then, the sorting method shouldn't matter as long as they're all in the correct order

Although it seems like invoking .Values generates a new list of values each time, but oh well. I can cache it I guess

That sucks. From this:

I thought you meant you had the onTrigger event working for you at some point.

The project page has some positive comments rather than a "Sorry it's broken, will fix after I get back from invading Poland." There are a few complaints about collision detections intermittently failing. Not sure if user error or lib.

Meh, it's a perfect storm of Unity lib and 1st time maintainer importing other libs, so you probably dodged a bullet, TBH.

That was my first reply… I just though by "getting accustomed" to having "everything you need" implied having used a working system at one point. It was a miscommunication as it seems they were still in the process of adapting their code to the platform.

Also, did you just assume my mortar?

I started doing some bare metal programming, that is, coding without OS, essentially making your own tiny OS. I want to do couple of simple games such as pong, tetris, breakout etc. Here are my first steps into that. Not a game yet, but a fake "loader" screen that is supposed to look like the C64 (Because why not). I include the program, you can run it in virtual box as a floppy image or even put it on your real floppy disk and boot from it. If your BIOS lets you chose what it emulates with USB pendrives, select the floppy disk and you can even boot from a USB.

Download it : www1.zippyshare.com/v/DRaeRGJB/file.html

Also with the linearized version of the quad/octree the position is intrinsically sorted due to the z-ordering.

So 2D I take it?
What's the purpose of the positions?
What features do you specifically need?

I forgot to mention that it's all in assembly. I will release the source when I finish it

I have a 256x256 grid with up to 2000 active objects on it. Specifically, I need to update the objects in a precise order, starting with the one in the lower right, working backwards towards the origin.

So does this mean a wall, or just anything that it can't go to (like a neighbor block with a higher pressure)?

Found it on CGPeers, it's freeleach:
cgpeers.com/torrents.php?id=47275&torrentid=47219#torrent47219

Look at z-ordering, it may be a match for what you're looking for; or at least it may give u some ideas.

Otherwise I'm not entirely sure, as a tree (specialized spatial data structure) is the best solution here, and using only a list is just a bad idea imo (u can do a list, and a tree f.e. -> tree sorts list, Which is way better).

Also… A spatial data structure only needs to be slightly updated when something moves.
So you're not re-sorting/iterating the entire list like u would otherwise for a sorting algorithm.
U could also get a specific order for iterating pretty easily with tiny overhead.

sorry for scattered replies, on mobile, and not feeling 100% atm

I don't see how this is hard, just use an array of 65536 pointers (256*256 pointers), it's only 2MB in size. The pointer's null when there's nothing, and you can keep a 2k item limit manually by incrementing and decrementing an integer.
You're not even close to any kind of memory limit unless you're developing for a literal calculator, and even for those I actually think the new ones could easily handle this.

Iterating 65000 items 60 times a frame is going to be very wasteful.

65000 if checks per frame is almost nothing at all, especially if the majority of it is null pointers like in this case due to any modern day CPU's branch prediction.

That is an incredibly wasteful way of doing it, you only have 2000 objects at most so, you only need to keep 2000 objects in memory, and keep 16 bits to say what the position is.

Really I would use a tree if you need to have them sorted, so just add an extra 16 bits onto your struct for the object, and a pointer, so this is 48 more bits * 2000, so this is 1.2 KB instead of 2 MB…

see how ridiculous this brute-force approach is?

What's ridiculous is caring about using 0.00000015% of the available RAM instead of 0.00003125%

Just walls and other immovable things, but only for a simple water sim.

You don't have to negate the vector. You can just stop the water, then add that velocity to the blocked item. If the item doesn't move, then pressure will build up against the immovable thing and naturally bounce back next physics tick as it tries to equalize.

That may be better if you have solid things which water can't penetrate but which can move freely. In that case you'll record the water's desire to move into the solid thing by attributing its velocity to the thing but not moving the water.

Note: This is how those desktop kinetic balls work. The one ball tries to move, but hits the stationary one. It stops but transfers its velocity (kinetic energy) to the next stationary ball. This repeats down the chain until the ball on the other and tries to move and actually can.

You can sum all the forces going into a solid object (a collection of cells that move in unison – store a pointer per cell that points to a meta object). Such solid objects are sort of a pressure gague. If the total forces acting on the object overcome its resting friction then the object can try to move (and moving the water directly in front of it as well).

I'd save hydraulic systems (water that moves rods) until after you have the general water simulation running. A piston is just a special case of a bunch of cells that move at once, but which nothing can move into.

Reflecting on impact would assumes that walls are things 100% immobile, they have infinite friction so the Newtonian physics equation would immediately apply 100% of the equal and opposite force back to the water cell. Reflecting is imperfect because it doesn't account for heat loss due to direction change (friction is your "heat loss").

Summing the velocity vectors into each cell as the water flows will take care of intersection of two opposing water flows. Opposing pressure waves will then naturally sum or cancel depending on if it's a positive pressure or vacuum being filled. The water doesn't see itself as a collier, the water count redistribution pass handles that.

You don't use your system RAM alone.. Its shared with everyone else. Being greedy doesn't work and will turn your game into a slow piece of shit notorious for eating up tons of resources. Imagine applying this line of Indian-tier thinking to every single other problem you will have to solve? This is why things like C# and Java exist, because of people who cant be trusted to implement the correct data structure.

Wait, are the objects keeping track of their hierarchy? That seems a little redundant

For the love of Horus, stop making fucking alien proportions, stop making unappealing anime girls from your imagination. Use a picture of an asian model. The Hugh Mungus issue that you guys don't want to see is that your weebfaggotry is inhibiting your artistic abilities. I know you like anime and this imageboard is based on some gook model and all that smug shit, but fuck, have some shame, don't be afraid to use reference. There's thousands of references online.

...

It can be a valid art style if done well.

The point is that you should have an understanding of how real anatomy works before doing stylization. Not "ugh fuck anime", which is a what a lot of art students misguidedly think when their teachers tell them to cut it out.

It's been a while since I used C#, but it's got a similar collections interface to many other langs.

So, what you want to do is implement the Comparable interface. Objects which express this interface (excuse my non native speak) can tell a container how they want to be sorted. So then you implement the comparison methods on your game objects. Your code will have an object look at itself and look at the other object.

If the other object is greater in Y (away from origin) then return a value indicating it sorts first (less). If the Y's are equal, then you check the X's and return (less) for greater values of X. Otherwise X and Y are the same and you return (equal).

This keep your objects in any sorted list collection and they will automatically be sorted.

The kicker is: Some collections may assume the sort order is static. So, you'll either have to call "list.sort()" manually (if there is such a method), or just remove the items and add them back to the list. If your list is backed by a tree then this won't be too taxing. If most objects are moving between grids each frame then just keep two containers open and add them from one container to the next each frame. 2000 items once is faster than 1500 items removed and then inserted again (because tree balancing / multi node iterations / etc.)

It's dangerous to go alone, here take this:
msdn.microsoft.com/en-us/library/system.icomparable.compareto(v=vs.110).aspx

I don't have to, because it is not relevant to this situation. In this situation, doing the lazy "wasteful" implementation still uses such an insignificant amount of resources that it doesn't matter. You only optimize things if they need to be optimized. And less than a millionth of the RAM and most likely less CPU cycles than his actual objects themselves will take is not something that needs to be optimized.

Wait, people using the wrong data structure caused managed languages to exist?

Agreed. I remember being a cringy little faggot in high school and drawing very ugly "anime girls" in the corner of the classroom. It wasnt until I started paying attention to skeleton sketching is when I started to improve. It's easy to get inspired by anime, but without basic anatomy skills, it's safe to say everything you make will be Chris-Chan tier trash. It's funny, thinking back, I remember thinking that I would be some sort of "free hand prodigy" if I just keep drawing my shitty OC character full of angst and pent up hormones.

You giganiggers, I'm not talking about the "dimension on which it is being depicted in", I'm talking about the big difference between the quality of work that would result from using and not using reference. I mean, look, would you rather fap to this or that?

I know user is not a pro artist, but it's I also know that is not very hard to use two pictures of a nude model, front and side, and place them on a plain inside blender or zbrush or 3dmax and sculpt that shit from the reference.

wow dude stop hating on anime you normalfag
go back to >>>/reddit/

There's already a SortedList that implements an IComparer behind the scenes, where T is the sorting type. Since my objects' position is simply a ushort, that works out. I thought that pulling and reinserting the object would be the way to go, too.

I don't have to do a sorting action on them every frame; that only happens when an object is finished moving. However, I do need them to be sorted, so I can iterate them in a predictable order.

No. Among other things, it was so you only needed to write code once, and not have to worry about manual memory management, or machine-specific code (although this still exists to a small extent, especially if you're doing things like making a low-level graphic library).

Although, C# did straight up copy Java; Microsoft wanted to poach existing Java programmers by making their language as functionally identical as possible (and to be fair, C# does run a lot better than Java does)

...

C# is slower than Java and is available on far fewer platforms.

It is a less painful language to program in though.

lose the hat and give her glasses and shorter hair.

Jesus Christ, that flat color with sharp vector cut angle art style always annoyed the hell out of me as a kid. It's one of the reasons why I hated Kim Possible as a kid too.

Let's make a game /agdg/

I'll be the ideas guy

I'll be the coffee girl/office cumdumpster.

Great! Here's your 70% cut

Every time I see a language that seems interesting, there's always some feature or convention that fouls it for me

Your advice is being trashed because you don't understand it yourself, you're repeating it for the wrong reasons, and you come off as a complete faggot while at it.

Point taken. A lot of people don't realize that Anime-style is what you get when you try to draw a 3D-ish girl onto 2D animation cells. The anatomy and what not is important, the face is usually rendered with reusable cells because you'll have a set of them already drawn / painted and a set of eyes and mouths and etc. to use with them. 2D anime is trying to look pretty 3D as cheap as possible. Then there are highly stylistic works, but those still follow 3D movement rules and then exaggerate.

Man DO NOT start asking "would you rather fap to __?" Some /d/evians fap to smart-homes, FFS.

What kind of ride are you trying to take me into?

Kill yourself as fast as you can for being a pedantic as fuck nigger, no seriously, kill yourself now, get hit by lightning so you turn into the flash and kill yourself straight into the speed force.

Just wish it user would spend at least 10 minutes to set up the proper workflow to success instead of winging it.

There's a difference between 'not specific enough' and 'mspaint strawman tier'.

I'm just advising user to be patient and use reference and study attractive proportions, which will result in appealing anime girls and not retarded aliens, like all those 3 pictures that you just posted.

I'll be the engine dev.

Check out this amazing render from my software rasterizer written in Java. It's got reflection mapping and everything, just haven't textured the model yet. There are also procedurally generated terrain and textures. It's 100% portable and no need to fight over OpenGL or DirectX – We'll use neither and have full control over the pixels since our engine code draws each one itself, without hardware acceleration.

Bonus, no need to learn how to use blender, you can use the 3D editor that's built into the engine as soon as I'm done coding it.

Well, got my objects roughly ordered, plus a rough outline for handling input and drawing and shit. Not sure how I'm going to implement object interactions or messaging, but I'll worry about that later.

That's enough work for today

Three more longsword attacks, a getting hit frame, and ladder climbing animation is all that's left before I have run out of excuses to decide on a platformer/metroidvaniaesque engine to actually make the game in.

I'm mostly retarded when it comes to high-level coding so I'm probably going to go with gamemaker.

Thanks again. I have the basics down (pressure, no velocity), but a lot of stability problems. It seems like two of them like to just start switching with each other, and never equalize. Any tips for this? Or is that something that velocity would somehow fix?

Kotamu.com (please use archive.is)m.au/2015/09/the-five-rules-of-australias-most-successful-game-creator/

Almost done, I will need to work on the sole a little because it doesn't give me a good shape but what you think? this is of course the base, so check topology, anything else can be just sculpted.

I've done some realism sculpting in the past, this was a few months ago. I think it's important to practice both anime and realistic sculpting. I'm going to start sculpting again because I'm out-of-touch with the basic shapes at least.

I don't think stuff like that are easily importable. Especially with all those branching out.
Although, if one were to make a standard for those, then there might be a time that those are easily importable.

Good job. Have some dubs , on the house.

Cool! The velocity will just help them slosh and splash better, and let you make fountains and such. The fluctuations are normal for that kind of sim. The overall shape should be mostly stable.

Just imagine any cell with > 0 being water colored. 1 and up are rendered similarly so the user won't see the tiny pressure changes fluctuating unless you just want to render it for them, and then you can have ranges for color of blue / sprite, etc. So that's a dampening factor.

If you need the values to be stable you can apply a viscosity factor to the fluid sim itself. That would mean rather than just flowing from higher to lower, you flow from higher to lower only if the difference exceeds a threshold. So, you could say, don't move from higher to lower unless the difference is > 3. That will reduce the vibrations but not eliminate them, unless you set the threshold at 8 or higher. Then instead of adding 1 water at a time to the sim you'd probably add more water to force it to flow better.

A higher flow threshold makes the water less viscous and more sticky.

The values fluctuating is kind of how real atoms vibrate anyway. Imagine adding dye to the water. That fluctuation stuff shows how water would disperse the dye.

Just download a human template with correct anatomy and start sculpting from there mate. Apart from the head and slight proportion changes and exaggerations, there isn't much difference between anime- and real bodies.
Honestly i would just go for Korean 2d/3d mix artstyle, in my opinion they've managed to find a very good balance.

Anatomy is very simple, this is a great tutorial; once you understand this it becomes pretty easy to apply it to any anime girl you want to make.

You dumb sack of shit. What the hell do you think 2D women are based on if not 3D women? You must be the stupidest motherfucker alive. You think because real women are imperfect that makes all 2D women superior? WRONG, dipshit. Just like 3D, there are varying degrees of actual beauty in 2D, but in 2D it is defined by artist skill rather than genes and chance. Anime is prone to some of the WORST depictions of women outside fucking modern art. If you honestly believe 2D is always inherently better just because it has a higher ceiling for perfection, you need to lock yourself in a walk in freezer.

Forgot to mention, you should probably try to move more than one water at a time.

If the difference is great enough move more waters, and if you're summing velocity then that means you're adding a multiple of the vector to indicate more flow.

That allows more pressure to push harder, and should propagate the water faster.

I worked on my game a little more scirra.com/arcade/action-games/first-game-18025?action=updated

I also set it up for multiplayer. I figure it'd be easier to do that now than run into any brick walls later on.

Now to actually add brick walls

Dubs and words of truth

Ever heard "She's a Brick House?" or "She's homely"? That's because 3D women are based on houses.

3D women look the way they do because of millions of years of selective breeding to look more like the idolized 2D cave paintings men prefer.

...

I'll help program as long as we use this user's engine.

of course you fags have to go and characterize shit. every fucking time with every single thing. this shit? its a trait of autism. visualizing a concept like game development as a person is autism. can't any of you retards take 5 steps back and just realize that? you do the same shit with video game consoles, boards, even just basic concepts like this. its borderline sonic OC tier.
to top it off, it looks terrible. you make this generic looking anime fag and try to make her look "THICC". the only retards who use this are either 14 y/os who watch samurai jack and go on funnyjunk, or are literal braindead niggers. looking past the terrible design, your art is fucking horrendous. it looks like something a tranny *hint hint john flynt* would draw and put in their calarts portfolio.
seriously, kindly never post here ever again you faggot. you make everyone who tries look bad.

medium.com/this-happened-to-me/10-ways-to-make-your-video-go-viral-d19d9b9465de

Autism is good.

Just try to ignore it and post on topic, making longwinded shitposts in response doesn't help anything.

...

...

...

...

Even if they were perfectly 2d it wouldn't make the majority of women not shit.

you know a lot about funnyjunk cancerous user base for someone who doesn't visit funnyjunk

Quick sculpting onto animu face and other stuff, maybe I need to find a better base myself IDK, doesn't look too good but it was a nice practice.

It looks like a tranny with plastic surgery. Nice job, user, keep practicing.

Need opinions on how to improve this shit.

I really need to make a base with way less polygons so I can really sculpt it with more simple controls, or so I think, is been weeks and I just want to make a new one from scratch.

i love how you went from a total faggot to someone that is (sorta) willing to learn

thanks.

is a prototype for mobile casual garbage.

is not a shmup.

not really, is more about a swarm and they become faster until you lose.

dunno, can work.

sure, need to add juice.

is for touchscreens.

In any case, thanks for your opinions.

here, watch this.

archive.is/RKOlI

It's an article trying to rationalize why 5 borderline absolute shit games of 7 have made it to #1. Let's not mention the insider connections. Let's not mention that iOS and Android are gamed as fuck.

Let's reference Flappy Birds with a straight face, and ignore the fact that the dev obviously purchased downloads. The game was shit, casuals have no fucking taste. They buy what they think other people like / are playing, and then pretend to enjoy it too, and then "journalists" even write articles about how "it has a certain something I can't put my finger on which makes it popular" – That something being paid for marketing buzz.

Seriously, everyone buys downloads in the app stores. It's not a secret how you get #1. It's $$$$. That's also how you get Arnold (((Schwartz)))d.Nigger to star in your 3D rendered commercial for a shitty 2D 4x. The app rankings are gamed. The walled gardeners do not care because they make as much or more money in people buying their own games as legit purchases / downloads. They farm aspiring devs who don't break even, and re-release a clone under a more heavily marketed name. Like there weren't a shit load of frogger clones.

"crossy road" purposefully invokes "flappy bird" because it's the same shitty situation:
bluecloudsolutions.com/blog/flappy-birds-smoke-mirrors-scamming-app-store/

It's called "marketing". Those 10 rules are 100% bullshit. "Make the game for the smallest possible demographic: One person." This shit reads like how exactly NOT to be successful.

Read it if you want to analyze how the propaganda machine paints itself with a veneer of legitimacy.

blackhatworld.com/seo/where-to-buy-app-downloads-reviews-ratings.653506/

big asses is a nigger fetish
are you a nigger user?

please make my dream video game for me please

Get on my level, I prefer 9th dimensional women. Low level faggot, pshhh

that face is disturbing…

why user, why

asses are great at almost any size as long as they match one's contours (flowing from thighs to the buttocks at a front/back angle, etc).
absurdly large asses generally do not fall under this category, so most, would not like it unless a particular positions makes the above occur.

upload your design doc and maybe if it's realistic some uncreative anons may steal some inspiration.

I'll give you these reasons:

It's an easy market, I showed you that video because I'm researching what's the secret to viral shit.

ah, you're in it for a quick buck…

There are two ways of getting people to download your shit in the Mobile market.
That's basically it.

I'm trying to learn how to make viral shit.
watching that video and it's like wow, is magical BS.

Going to make a clicker game about dogs versus cats, the doggos have anime dresses (microtransactions) and they use real fast food as advertisement.

going for maximum doritos cancer.

triple 64 checked

Your game is perfect the way it is. Just put some lol random faces on the squares that change before disappearing.

Rip off a background from some well known game, like Mario or Sonic, then purchase some downloads. It will also help if you create a twitter account and hate on "ebil gamergate", and claim they're harassing you out of games for being a self identified stronk wymmin.

not even trolling

3.
Buy downloads like every "viral" app ever.

Viral is code for "It's totally orgaic for this shit to have so many downloads and ratings, 25000% more ratings per install than average is normal, people just love this shit for some weird reason."

go viral = get AIDS.

not sure what else i was expecting. incredible.
i guess solo really is the only way to go.

...

well that fucken sucks, how much time did you invest?
1ma is really the only way you can guarantee you won't get fucked over, and also having creative control is nice.
Though there's obvious tradeoffs as you've realized.
Hey maybe someone here would want to teamup, but obviously you've got to have skills to offer.
List em off and u might find you dream dev partner.

that's not really a good enough reason.
Basically, you don't give a shit about games, and only care about marketing tactics.
You're like a born again shill trying to find their new niche; fuck off goldstein.

I like games, but I also do consider this a job.

also marketing is not bad or good, is just a skill to have.

good games gets released by the millions, what will set me apart?

Actually if you know yourself then 1ma is the least reliable method. At least for me.

Terrible excuse.
Even if you're performing a job there's an extent at which one with good characteristics can bend their principles before they break.

Of course, any forward thinking dev has researched this topic thoroughly.
However, you're making this the point of your project; which ascribes certain qualities to your motivations.
Your previous posts state that you're willingly entering the mobileware market, and are attempting to have your "game" that's analogous to a shit in a designated street stepped on; which will probably lead to these practices being passed on to your jewish friends (yeah I know how u fucks operate) to make additional shitty games on other platforms (remember it's a job, so you'll pursue a position if you become good at marketing, morals? fuck that, pass this shit on).
Can you see why someone who wants those millions of good games to be noticed, is annoyed by you, and people of your ilk?
You're contributing to the problem of shit games piling up, and I certainly cannot respect that.
Also, any self-respecting dev who does this for making one of those good games - that you're making harder to find - that they can be proud of should look on you with disgust.

The answer to that question is surprisingly simple:
however, I'm happy to know you'll fail in this regard :^)

depends on who that is, and their personality.

some people work better in groups, and some work better alone.
I'm in the latter category most of the time, but if the person is as dedicated, and passionate as I am; then the time requirement becomes less of a factor to attaining (what would be) "our" standard of quality.

there's nothing wrong with learning about marketing and making viral content.

you seem like another typical pol tryhard.

It's 380d39, but my IP changed.

more than that, people finding managing things on the heap too hard… it goes on. Those languages implement all the data structures for you. Java does not trust the programmer with pointers because it is made for Indians like 3b57c0.

Tree is just a sorting method, you can sort things in other ways if you wanted too.

writing good code is about doing things the correct way, not by brute forcing. In this case you are talking about 1666x less memory. Its not even a more complicated thing to do, since the code isn't doing all the redundant shit you are doing. Also, it has a complexity that will go up exponentially. Lets say I want to have a 512x512 grid? Now I have to take 2MB^2, 4MB… then 8MB for 1024, then if I want 2048x2048 I will take 64 MB. Meanwhile the tree I suggested will increase in complexity at a hugely slow rate… so if I want to say represent a bigger grid I only need to add an extra two bytes per object, so and suddenly… I can represent 16385x16385 grid… in 24 KB… meanwhile if its done your way the data structure will take.. 1 GB…

Its obvious which one is the better approach and WHY brute force, exponentially expanding algorithims are bad… and the fact that I have to argue this is amazing.

you sound like one of those bugchasers, faggot.
DESIGNATED. SHITTING. PAJEET.
kek

Having an issue with unity, and the only thing I can find online is the unity team screeching at people acting for help. I'm using a SyncStructList, and it syncs new items when I ADD to it, but doesn't initially sync it properly.

Anyone run into this before? Documentation is shit, the forums are shit.

Naturally, if the client reconnects, they just get a blue screen until the server starts spawning event more new blocks. I just want the list to fucking sync when it's being sent over, as promised

Shit like this really makes me appreciate python.

Learn to program you subhuman.

Sorry, have you tried writing a self-allocator with variable data inputs? I'm just one man, go show me your compiler you nigger.

I don't program much. But I don't need to lay an egg to be able to smell a bad one, and good C code does not do that. There are useful tools to track down bad memory usage though like valgrind, or whatever debugger you have on hand.

1) Assess if C is the best choice for whatever it is you're doing
2) If it is, quit your bitching
3) Use assertions in your debug build
4) Learn how to use a fucking debugger PROPERLY

it sounds like he's learning things right now, probably the best idea is to let him figure it out

The way the code is structured gbd can't really run on it. gdb is really only for static programs.

OSdev is not a fucking easy thing to do, user.

yeah, I didn't say that, I was pointing out that it didn't seem like you were asking for help, but that you were figuring things out by yourself

Aight, well what are you developing then?

youtube.com/watch?v=AyBfhSWhmdc&app=desktop

I'm working on a 2d sprites in 3d engine stealth platformer with Xenogears-esque combat. So far I have doom-like sprite rotation working, got a parser that I am heavily modifying to assign textures to actors. Next I need a sprite wrangler and sprite manager wrangler, then I have to write c++ code to create a sprite animation registry. In the meantime, here is Fei Fong Wong. I ripped his standing, walking, running, and jumping anims. THANKS DEBUG ROOM :D :D ;D

I'm a fucking idiot, I should point out that this is being done in Unreal Engine 1 (Unreal Gold 227i).

Christ this is annoying

Here's the source, if anyone cares to take a look. I haven't exactly refactored it, it's barely working, and I'm giving up on it for tonight.

hastebin.com/cupuhogito.cpp

I-I recognize that butt

Try decreasing temperature of those red hot blocks, maybe it's just boiling.

only two hundred drawings to go

A single for loop with one if check and then his object-specific codes. Zero management needed, zero sorting needed, almost zero code needed.
1666x less than basically nothing is basically nothing gained.
But he doesn't, that's the point. You keep talking about hypothetical situations where an optimization would be needed, which will never happen in reality because this guy's design has been set in stone since before he began working on it.

You really, REALLY underestimate how hard it is to make money on that market. You're competing with companies that have been doing this for years and have entire teams dedicated to study the retardation of normalfags to best exploit it.

You're not competing for normalfag money as much as you are competing for normalfag time and attention.

Whatever it is you're trying to do with C, I'm certain my answer would be: "Yep, I've done it." or "Kek, no, but here's how you should do it."

Here's some debugging output from my memory manager dumping the page-level record keeping structures. I replace malloc() and also provide a thread local allocation. I don't use one allocator - I have a hierarchy of allocators for different situations (one of those situations being automatic garbage collection). The top level memory manger manages the memory that allocators use and shuffles pages around between thread-local allocators if needed rather than making an sbrk() call to increase the .data segment.

Yeah, the allocator allocates and manages itself. I've used it in OS development. Replace "threads" above with "processes" and bob's your uncle.

I have a macro for new() in C:
MyClass * instance = new( MyClass( someVar ) );

This macro uses ## to glue "AllocAttrib__" to the the front of the constructor passed in, resulting in:
MyClass * instance = Allocinline( AllocAttribMyClass( … ), MyClass( … ) )

AllocAttrib__Xxx is a macro that accepts any number of parameters but only returns a static structure which include a sizeof( MyClass ) property - this macro was generated earlier when the class was declared via another macro:
newType( class MyClass { … } );

The "new" code is finally transformed into:
MyClass * instance = neAlloc( neMyClass_attrib );
ne__MyClass_Constructor( instance, … );

It looks convoluted, but it's actually simpler and faster than the C++ allocator interface.

When I declare a new type I just specify which type of allocator it wants to use. That behavior can also be overridden / modified.

At the top you see some "Chain" output. That's a hybrid array list / linked list from my hybrid OOP / scripting language which can co-exist with C structures in the same allocator system, and has data and objects with methods that can be called directly through C
->dyn-> is a "virtual" function call for polymorphism in C.

I'm able to inspect such things at runtime because I implemented runtime type reflection in C.

Most serious game projects have implemented a custom allocation and memory management system. I even have an optional generational garbage collector as one of the allocation strategies a type or allocation call can request. In the old days every gamedev was an OS dev, since we had no OS just raw hardware.

Try not to pretend like you're doing shit no one will understand. People have done it before, and probably have good solutions or advice if you ask smartly. My advice would be to develop your OS more modularly. Make the memory management systems to be stand alone so it can be reused in non-OS applications. Treat threads as processes. Since you're not Use memcheck / valgrind to find the cause of your segfaults and assert. On windows the memory debugging is shittier (still can't find something to replace cachegrind on Lin) , but there are tools like Dr. Memory, VMMap, AMD Code Analyst, verysleepy, etc.

Yep, done it for decades. Here additional pic relateds of a 16 bit OS for gamedev on Intel 80186/7 written in assembly. The kernel has an interactive mode for manual commands and a built in debugger for viewing program state and editing memory structures. Here I show allocating and deallocating memory, then open the memory editor and show the record keeping structures.

Build such a memory debugging tool 1st, make it able to be switched.

Maybe later, when you're being rational. You realize you're being the nigger here, right? People are just trying to help. You're chimping out about how hard your unique struggle is living in fertile but undeveloped land, compared to whitie with his high level language and pre-made OS. Your problems have solutions. We've been through the shit too, and might have some sage advice. If you don't want help, man up and quit yer bitchin.

Here's to hoping that code tags still work here.
MyClass * instance = new( MyClass( someVar ) );This macro uses ## to glue "AllocAttrib__" to the the front of the constructor passed in, resulting in:MyClass * instance = Alloc__inline( AllocAttrib__MyClass( … ), MyClass( … ) )AllocAttrib__Xxx is a macro that accepts any number of parameters but only returns a static structure which include a sizeof( MyClass ) property - this macro was generated earlier when the class was declared via another macro:newType( class MyClass { … } );The "new" code is finally transformed into:MyClass * instance = ne__Alloc( ne__MyClass_attrib );ne__MyClass_Constructor( instance, … );

oops:
*Then, since your're not running the allocator system in an OS mode, and can run it in an application: Use memcheck …

That lets you debug your allocator from inside an existing OS before being deployed in your OS.

Thanks, that was lazy of me. I doubt it will matter much.

Eventually I'll release my source code. It's got some rare features enginedevs might like to checkout. Part of my gamedev motivaiton is to disentangle my code with work related code so I can release it w/o copyright worries.

this is complete bullshit
you should always aim to write code as simple as you possibly can and only consider performance if you know for a fact that it will have a notable impact on your game.
a tree is definitely more complicated than an array especially if you haven't worked with it before.

Ah yes, the legendary Toothbrush of Samus.

This is where I give up gamedev for months again, isn't it

Birds dont have teeth

or do they

I started a new mesh. I went as far as sort of tracing the outlines of the models because I'm a scrub. but I didn't copy any meshes or loops so I guess it's alright? It's for practice anyway.

yeah, I've seen a lot of artists in sketchfab and pinterest that happen to be korean. I'll probably try to do a realistic body model and grind sculpting from there when I have the chance.

Such is the life of the Architect and his simulated matrix of reality.

I don't have time to test the sauce right now. But that one water hopping up and down on that platform shouldn't be, also it should be deciding to go left or right sometimes. That's probably where the sim needs adjusting. You should prefer moving down if possible, then left or right if possible, then up as the last resort.

Guys I broke unity what do

Return to the Source!
Kiss Trinity goodbye, Neo.

That's what you get for fugging that Unity slut.

Looks pretty rad actually, like he is running so fast he is transcending through time and dimensions.

...

I think it'll smooth out if you make him run faster.

Well, I guess he was right.

I agree with you in principal, but in this specific case the dev is using C#, which has nice interfaces to trees, like SortedList. So they should just use the SortedList collection API, and implement IComparable on the objects that need to be sorted to provide a custom sort order. In this case all you literally have to do is write a function that compares two objects based on their location and returns a value indicating which one goes first. Done.

Then you just treat the objects like regular and put them in a SortedList, and don't worry about making some custom pointer array optimization system.

This is the simplest answer, and will out perform iterating across a sparsely populated array, especially a big one, because SortedList is backed by a tree internally, but you just have to be familiar with a regular old list.

If you're also trying to get near constant time lookup by x,y location, use XY as a key in a key/value store collection that's backed by a hash table.

TL;DR: Just tell C# how you want to order the objects, that's really it. The answer is the same in Java, C++ and C (in C: qsort, heapsort, mergesort allow you to provide your custom ordering function).

Who do you think we are? Some kind of nodevs?

Don't believe in yourself. Believe in the AGDG who believes in you!

Stick with it. Eventually you'll converge on a solution, and then you'll have it solved in all future universes too.

KAMINA
GLASSES
ARE
RED
NOT
ORANGE

Centralist actually, I just like to make myself laugh.

They are red. The glasses only appear orange-red because of the composition.

I went with SortedList. It has indexing so I dont have to fuck with a collection being built to get what I want. Its a little slower on add or remove than SortedDictionary, but I'm going to be iterating it more than adding or removing elements. (I guess I build a secondary list of elements to pull then readd after the iteration so traversal doesnt get fucked up)

Even though my map is 256x256, positions are expressed as ushort, so I dont need to implement a custom struct or comparer but its nice to have the option

You make terrible posts.

...

you have only yourself to blame tbh fam

fag.

(You)

I think I'll scrap weapons having down attacks. Standing attacks and jump attacks seem to be adequate.

Ladder climbing, jumpstab, and upstab are all that's left.

Train of thought: Since the only air attacks people used in zelda 2 and shovel knight were the downstab, why not have that be the air attack?

looking pretty good

Is there a better way to do ground clamping than just applying a down vector to my character?

Like, that sounds very simplistic, is that a problem?

using unity

Updated race descriptions and features of Humans, Fae and Dracari
fablesoflaetus.com/lore2/

...

It looks better in the blog format
fablesoflaetus.com/

Only reason I can think why you would check your dubs and not have that be the air attack would be for consistency: the principal of least surprise. Attack while standing does X. Say you press attack while you're falling down a little 8px step. Suddenly you're doing a ground attack instead of fighting while walking down stairs.

You could try to indicate to the player somehow that when you're X units above the ground then you no longer do a regular attack, it's now a ground attack. But then you have to figure out how close is too close to the ground to do a ground attack? For the sake of simplicity it's less complex to just always do standing attacks while in the air / falling unless the player is explicitly making it a different attack by pressing up / dn while in the air.

Then you just have to detect whether you're on the ground or not before initiating a ground attack, which you should already know from the last time you stopped falling down.

stop falling down!

I always liked "cancels". Which are basically just shortcuts. For instance, a ground attack while getting knocked back could cancel into the standing posture. Still being knocked back, but able to deliver standing attack, and being ready walk again sooner.

In SMB3 with the Racoon or Tanooki suit, there's a slight bug where you can fly while crouched, allowing you to fit easier into certain places.

I used to like to cancel the BFG9000's audio event by pressing the fire button right as I hit the ground, making the "umf" sound silence that loud charge up sound which would otherwise make people flee… but that was exploiting the limited voices of the audio system.

If the creators had thought of some of those eventualities they might have made mechanics out of them – maybe made air-crouching part of flying, or a way to muffle certain weapons, etc.

That said, most people don't micromanage movements when playing platformers and FPSs, but some fighting game players just can't help it.

If you can discover those little opportunities to add just a little extra control then they're like much appreciated easter eggs for hardcore autistic fans, and at the least they won't become bugs / glitches.

user please download a decent theme made in this decade

Sorry i have zero experience in website layout, i will fix it later, gotta prepare the content for tomorrow's lore post first

going to add visuals & sfx to my pong-breakout clone and upload it to glorioistrainwrecks

fuck I had a full knight of sleep

More X11 problems:

AFAIK this means that unless I figure out a workaround, X11 wants to be able to block your program when you aren't receiving messages very badly, which is really the opposite of what I want to do since I am making a game, not a spreadsheet, so only executing at X11's discretion is an obviously bad idea…

Its clear that this was not made with games in mind, im hoping to figure this out soon.

X11 was built to run asynchronously in a client / server model.

Most event systems, QT, Java, Javascript, etc. deploy a custom event thread separate from main execution. That's generally the way I do things – have an event IO thread, and at that barrier I also platform agnostic application-level events to feed the rest of the program.

However, X11 you don't have to employ multi-threading if you really don't want to, because it's a network transparent protocol that uses sockets. So, you can get at the socket yourself via ConnectionNumber().

// Get socket / fd / handle of X11 display for selects.
xfd = ConnectionNumber( display );

Then, just like any other non-blocking socket IO, use the select call to determine when that FD is ready. That's when you call your XNextEvent().

#include #include #include #include Display *dis;Window win;int x11_fd;fd_set in_fds;struct timeval tv;XEvent ev;int main() { dis = XOpenDisplay(NULL); win = XCreateSimpleWindow(dis, RootWindow(dis, 0), 1, 1, 256, 256, \ 0, BlackPixel (dis, 0), BlackPixel(dis, 0)); // You don't need all of these. Make the mask as you normally would. XSelectInput(dis, win, ExposureMask | KeyPressMask | KeyReleaseMask | PointerMotionMask | ButtonPressMask | ButtonReleaseMask | StructureNotifyMask ); XMapWindow(dis, win); XFlush(dis); // This returns the FD of the X11 display (or something like that) x11_fd = ConnectionNumber(dis); // Main loop while(1) { // Create a File Description Set containing x11_fd FD_ZERO(&in_fds); FD_SET(x11_fd, &in_fds); // Set our timer. One second sounds good. tv.tv_usec = 0; tv.tv_sec = 1; // Wait for X Event or a Timer int num_ready_fds = select(x11_fd + 1, &in_fds, NULL, NULL, &tv); if (num_ready_fds > 0) printf("Event Received!\n"); else if (num_ready_fds == 0) // Handle timer here printf("Timer Fired!\n"); else printf("An error occured!\n"); // Handle XEvents and flush the input while(XPending(dis)) XNextEvent(dis, &ev); } return(0);}

But remember: You could also just have an event thread that deals with the input system.

I should say, that 1s timeout can be reduced significantly, and pselect will let you specify nanosecond precision.

man select 2

saw this sample code on stackoverflow but I thought.. no way, X11 must have this feature in a function so you don't have to gut it like this to get the messages.

I DO want to multi-thread this, I just didn't think that I would be doing so NOW. I suppose there is a first time for everything…

Yeah, that example is one from a common set.

The reason that X11 doesn't have special features for nonblocking IO is that nonblocking IO and really fast OS level zero-copy memory mapped IPC are done via select calls… Creating a wrapper for select() would be just inefficient.

On windows you should also get familiar with select… Don't shrink away thinking it's slow on Linux like it is on MS Win.

A FD on Unix is one of the fastest ways to move data around in a computer.

Also, if you do much file IO and networking, just add those file and socked descriptors to the fd_set (in_fds) here. Then iterate the fd_set to check which ones have IO ready.

You've been a huge help. I think I'm pretty much done with the 'core' of it, minus some tweaking. I just need to figure out and implement a way to slowly stabilize it, so the water doesn't bounce around forever.

I see you too are a man of patrician taste

I just state the truth

What you guys think of my elemental icons so far?

Earth and water are good. Fire starts blending into the orb too much and air is fucking trash, no contrast at all I can barely see the symbol part of it

Where's heart?

But you are not supposed to see much in the air icon, shit is transparent :^)
The icon is the whole thing, not just the symbol in the center, in fact i was thinking about merging the water one more in the background

Died in the streets of Sao Paulo after stealing from a gang member

...

I'll get too it once I can get Xlib to stop crashing when I call the XCloseDisplay functions

First two looked like vulvas at first

I'm just going to use pixelart and discard the transparent pixels in the shader for the rest of my life, and never use semitransparent anything.

Transparency dither

if you are using the fixed function pipeline this is done with GL_ALPHA_TEST and the semitransparent level is set with GL_ALPHA_FUNC. I doubt this is what you are looking for though.

ok remade water and the last two common elements

now to make the 12 composite elements…

According to gdb it crashes outside of the main function?

Really its hard to tell what is happening because of how unreliable xlib and GLX are. The code works fine on windows and has no crashes, so i'm not worried about what i'm doing on the heap… I guess its time to install valgrind and figure out how that works.

First composite element:
Fire+Earth = Metal
I can't decide which symbol matches metal more, since this is an element used by smiths should i make it a hammer?

either a hammer or a gear works.

Or maybe a bucker for the radial symmetry…

I want to make a game in RPGmaker, and I've messed around with the program several times, made some little demos, etc. but I don't think I really know the specifics about balancing, or what items I should or shouldn't have, or what damage formulas I should use, really in-depth mechanical stuff like that.

Does anyone know of any good tutorials or sources dealing specifically with the general balancing and numbers of an RPG?

A gear is a nice idea too, lets see if i can make it look good

Play tons of JRPGs, thats how you gain experience
I suggest Breath of Fire 3 and Chrono Trigger, these are masterpieces
Also i don't remember RPGMaker letting you change damage formulas… but well the last time i used one was almost a decade ago

For stuff like that it's really exclusively just knowing and feeling what's right.
Not to mention, for balancing and shit like that you have to wait until you actually have to do some playtesting. That means after you've built all the base mechanics.

Just start making something, you'll figure it out as you go

We had discussion on this a while back.
Basically, the numbers are whatever you want them to be. Two related series: look at Mario and Luigi versus Paper Mario, for example. Both are about as difficult (read: piss easy). Yet, to me, Mario and Luigi seem way more powerful, endgame, than Paper Mario and his partners.

The average attack in Paper Mario is 10 hit points at high levels (13 charged up, and a chain of 10-9-8-7… as a multibounce or whatever).
In Mario and Luigi, though, even your most basic attacks will do upwards of 50 damage, with specials going for several hundred.
Then, of course, there's Final Fantasy and the like, which goes into the thousands, easily.

The numbers are, really, arbitrary, but bigger numbers give you the illusion of hitting harder.

As for balance, well, play the game.
A lot.
If, after a boss, you need to spend two hours grinding just to stand a chance against the enemies of the next area, your RPG is shit and padded to hell with boring grinding.

So… Just as good as all the other RPGs?

I hardly ever use switch statements. They're kind of shitty because they can only match exact cases (eg, no X > Y ranges). Also, if you match a number against an enumerated type, say a Direction enum that defines up, down, left, right, and none, there's nothing stopping the user from passing ((Direction)215) into the function and breaking it.

Yes, you can put in a default case, but that's only a bandage to bad design

Metal done
next is Fire+Air = Thunder

alright, thanks for the input

The white outline around the gear looks very off-putting. At a glance, it's like someone badly cut out a gear out of an image, didn't do a very good job and and left some white crud on it.
It makes the gear look fuzzy and soft, losing the sharpness of the teeth. Whatever subtlety there was in the contrast between the gear and the orb is lost.

Thunder done thats right, i said thunder, not lightning
Next up
Fire+Light = Ray

Hum, now that you mention, maybe i overdid it
Here the new and old one for comparison, what do you think?

Nevermind, did an even better one

Do other programs segfault when they close? Run them from terminal to see.

If other programs aren't segfaulting on close then it's probably not Xlib's fault.

I've had a crashy X, but usually that's when I'm using unstable builds of a distro, or or nightly builds even. It's rare on stable releases, unless you've got flakey hardware or driver issues.

I've got a whole folder of exploits that crash various parts of MS Win and Linux/BSD by doing bad things with Win32 / X11 event systems… X fixes theirs pretty quickly [hell, I do it myself]. Some MS Win bugs I reported several years ago are still viable today, they go slow unless the exploit is in the wild.

Try adding your window closing code to this small example program:

That'll help you separate X11 from GLX and pair down the problem. That code doesn't handle window close events though.
// Global scope.Atom killwindow;...// In your init// Allow interception of window close events.killwindow = XInternAtom( display, "WM_DELETE_WINDOW", False );XSetWMProtocols( display, window, &killwindow, 1 ); ...// example event loop:XEvent event;while ( XPending(display) ) { XNextEvent( display, &event ); if ( event.type == ClientMessage ) { if ( (Atom)event.xclient.data.l[0] == killwindow ) { /* Window closed. Call your clean up code & exit. */ } }}

Remember to destroy the context before the window.

// Remove the window from the context before destroying context and window.glXMakeCurrent( display, None, NULL );glXDestroyContext( display, context );XDestroyWindow( display, window );

And that is a very orange explosion.

If nothing else, it reads a lot better. The middle's a bit of an amorphous blob, with mainly just a little difference in color and little in value, but the teeth do at least have some contrast with the darker edges of the brown.
Maybe get the rest of them done. You might find some unifying design elements. Like being able to reuse earth. That has a nice look to it. Right now, the gear and explosion are the odd ones out. The rest generally have this shading thing going on, but these two just have disparate entities in them.

The air attack's startup hitbox hits overhead and in front. You'd suffer a bit of unexpected lag time after the air attack that would be compensated for by longer histun and more knockback against whatever was hit.

Also, the air and ground controls are supposed to feel completely different. The jumping controls are similar to CV4, while the ground controls are closer to battletoads. This is a deliberate thing. You have less control and fewer options in the air. You can't even turn around in the air.

I know for a fact that I am correctly processing ClientMessage's. However the issue is destroying the context

I call these functions in this order to shut down X and openGL.
...glXMakeCurrent(sys->disp, None, NULL);glXDestroyContext(sys->disp, sys->glc);...//this is where crashes start happening:XDestroyWindow(sys->disp, sys->wnd);XCloseDisplay(sys->disp); ...

Sometimes it crashes on XCloseDisplay. Other times, it crashes outside of main().
I have tried to remove XCloseDisplay but the crashes still occur. Sometimes this shows up:
*** Error in `./sigma2.bin': free(): invalid next size (normal): 0x09f7ae28 ***======= Backtrace: =========/lib/libc.so.6(+0x71030)[0xf7387030]/lib/libc.so.6(+0x7b833)[0xf7391833]/lib/libc.so.6(cfree+0x58)[0xf7396398]/usr/lib/xorg/modules/dri/i965_dri.so(+0x3371c3)[0xf6d011c3]/usr/lib/xorg/modules/dri/i965_dri.so(+0x3da007)[0xf6da4007]/usr/lib/xorg/modules/dri/i965_dri.so(+0x333dd6)[0xf6cfddd6]/usr/lib/libGL.so.1(+0x40a56)[0xf773ea56]/usr/lib/libGL.so.1(+0x17f33)[0xf7715f33]/usr/lib/libGL.so.1(+0x17fb2)[0xf7715fb2]/usr/lib/libGL.so.1(+0x180f0)[0xf77160f0]/usr/lib/libX11.so.6(XCloseDisplay+0xb3)[0xf74f7223]./sigma2.bin[0x804d28c]./sigma2.bin[0x8049a7e]/lib/libc.so.6(__libc_start_main+0x107)[0xf732e697]./sigma2.bin[0x8049b12]

Other times it doesnt. I havent run gdb with XCloseDisplay removed yet.

Okay, i removed XCloseDisplay and now it doesn't segfault at shutdown. I don't know what X is really doing but im hoping this behavior continues. I'll continue my normal work and see if it starts misbehaving again.

The effects of the combined elements are very specilized, i want them to contrast with the common elements
The colors gradients of the combined elements are based on the common elements they come from, if you zoom in there is some green from the air element on the center of the explosion
Ray element done
next Fire+Dark = Ash

I don't know why the fuck I keep having problems accessing other scripts but half of the times I can't seem to figure it out

private CharacterControllerUtility utility;void Start () { utility = GetComponent ();}void Update() { utility.ClampToGround ();}

Is this wrong?

alright, so this is whats happening:

most times it starts up and shuts down correctly. Sometimes it segfaults on startup and sometimes it segfaults on shutdown.

I don't know if this is drivers, but there is definitely some undefined behavior that X is doing and because of how sparse the documentation is I cant attribute it to anything in particular right now. i'll keep hunting with gdb.

If problems return then it sounds like a job for valgrind.

The –gen-suppressions=yes option is a useful tip.

Since the OS reclaims unfreed memory on exit, and X11 allocates a bunch of static memory that are used throughout the program, it doesn't free those before exiting. It would just be having the memory manager shuffle around values in the heap only to have the OS reclaim those pages a moment later. memcheck will warn about this situation of unfreed memory, so you can use a suppression file to suppress those notifications.

Hopefully if the memory it's using gets overwritten somewhere it could cause segfaults in library code even when it's not the library's fault. Memcheck can put memory guards around allocations to catch overruns as they hit the surrounding memory protections.

Intermittent error reports smell vaguely of memory bounds violations. If you do catch X doing something crazy, devs won't care about the report unless you've compiled it from sources. Bug chasers might. A distro that builds from source comes in handy here – time to unironically install gentoo. or Arch. Then you can step into the code and see exactly where something crashed. Of course if that code is running on bad data, then it's GIGO.

IMO, the #1 reason I port to Linux is just so I can use it's memory analysis tools. Supporting FLOSS users is just a side effect, really.

Phew, this one was a bit harder to make
Ash element allows casters to crate dark embers that act as false souls, basically elemental version of necromancy
Now
Water+Earth = Flora

Where have I heard this setting before?

(((pure coincidence)))

A FUCKING LEAF
Ok next is Water+Air = Squall

Tomorrow I'm gonna do it, I'm going to make game

the leaf one is actually nice

-ur bully

READ THE FUCKING OP, DO IT NOW

The original one? I missed you :3c

I used valgrind and solved my problem, its almost too embarrassing to post here.

Bad code:
XRRScreenSize *xss = XRRSizes(sys->disp,sys->screen,(int*)&sys->vidmode_count); sys->vidmode_list = calloc(1, sizeof(vidmode_t));

The significance of this is that every time I told X what the width and height I was using, this was a buffer overflow. So it was randomly crashing everywhere because I was stupid and assumed that my code was OK, and didn't bother looking.

Not getting any crashes anymore, i'm going to continue work now.

Squall done, got this effect accidentally but it worked quite well, i couldn't figure out what central icon i should add, so this saved my ass
(it was supposed to be weather related)

Water+Light = Prism
The boipussy healer element

PANIC
going from AGD to PRO in a week, with employees and salary from the 1st. How do I not sperge or let my alcoholism hurt my team-mates?

...

Do not drink alcohol?
Anyways good luck user!

Ok this one took some time to figure out, but kinda looks good i think

What do you guys think of this? Does it sound "gamey"?
soundcloud.com/graz-hero/boss
I threw it together in a few hours and it came out sounding sorta like Chrono Cross.
also does anyone need music for their game?

Sounds gamey, but a bit over-bearing with those repeted intro notes. Maybe adjust their pitch and have their volume lowered after the first few, to focus on the other instruments.

Could use some more bass too.

Ask again in the next thread, 2 pages to go.

I agree. I wanted to develop the intro more, but I just settled for a copy-paste since I was lazy.

you're only allowed to do that for art, code, and a game's design. And all the content.

I'd ask for music if I were a few months ahead in progress

It reminds me of Threads of Fate, but definitely a slower-paced bossfight like an RPG or something.

>also does anyone need music for their game?
I mean, the Souls series doesn't have music outside of bossrooms, and I'm gonna need the music made in famitracker or sound like it was made in it, but if you're willing to compose two dozen mostly ambient chiptune pieces and/or transpose a bunch of existing songs to famitracker I'd be glad to pull you on board.

I'll keep that in mind

I like Famitracker since I don't have to worry about having a good soundfont, but I don't think ambient is really my strong suit. Thanks for the offer though.

The first part is a bit repetitive, but very good, the second part is a bit too loud, but its good overall
Can you do comfy stuff? Like songs for a rune factory/harvest moon like game?

Don't be too embarrassed. We all do that.

I do the same thing every time I have a bug.

Sometimes I'm actually right and it is the library. Then I get to gloat (or tuck away a handy exploit). Most times I'm wrong, but it's worth questioning other people's code. That's how you find exploit vectors… for which there is a market, if not a bounty.

Do you think you are able to make things akin to a Metroid series song? Because that's the kind of ambience aimed at here. Something unrepetitive and fairly long.

Sure. But the more I think about it, the more I realize I really don't have any business offering to make music for people since I don't have a whole lot to show for my work, and I don't have any professional music education, sorry. But the night's still young, so maybe I'll put together a "comfy" piece and post it in the next thread.

I can't say I've tried making something like that unfortunately.

Oh well 5 elements to go and then i can post all my autistic ramblings about magic and shit
Goodnight you lovely faggots

Who the fuck cares? if you make good shit i will buy, but if you are like the last artist i hired to make music which i had to babysit, because if i didn't check on him every time he changed something he would completely ruin the song then no thanks.
It was expensive as fuck too, im still mad

Yeah, but without education, it's hard to tell if every decent thing I make isn't just a fluke. Like, I could convince myself I know how to make music, and then be completely stumped when I actually have an obligation to make it for someone. I mean clearly I have trouble being objective about it; repeating the same bar 4 times is retarded and I still convinced myself it was good enough to upload.

m8 i don't know how to draw, never did animations before, and don't know how to code but i'm still making this thing

Almost every time I post something like that, the response is something about how great RUST is or how you should just like use shared_ptr or how java is great

Not here, but in other places…

I was trying out musl-gcc but it turns out that to actually use it I would need to link everything else to musl and compile it all over again.. so its way too much of a pain to use C11 threads.h on linux. I'm going to just use pthreads since that is obviously easier and better than trying to find something standard compliant, unless there is some other way of doing it. Maybe if I were only on windows since Pelles C seems to be doing the best job in this department.

Jesus christ, I'm glad to avoid that mess altogether. Thank god good engines and free frameworks exist

Solved this problem:
XPending is the function I was looking for. Now I can do this in one thread without any problems.

The problem I am having with multi-threading I am having is that X is not responding to calls from glXSwapBuffers from thread 2, which is doing the graphics, whether I am using a message queue that involves blocking or not. It was easy to split the program into two threads.

Decided to use one thread, since this isnt a question that I want to take the time to answer. Its not a problem having it in the same thread, just because sys_update is so insignificant in its complexity and urgency.

I should stop spamming the thread with these posts

If people recall, I'm the guy that said he is pretty good with java already and wanted to git gud. I got some neat tips in regards to c# and C++. Would some kind anons give me the links and descriptions they gave last time again?

archive.8ch.net/posts.html#!v/12618051_#12625638

I did not know that we had working archives these days. We probably had them for a while now.
Thank you kindly, user. Sorry for the bother.

New bread?

STOP ASKING FOR NEW THREADS WHEN WERE ON PAGE 12
EVERY FUCKING TIME

we were in page 13 a minute ago dude chill

that's not how it works

im taking a screenshot next time then
im not crazy, im sure i saw it

...

Haha i was not crazy!

Assuming I'm a cheapskate and I don't want to spend money on an engine. Make my own from scratch(I've done it before in >java and it's a pain in the ass and the game was just ass) or are there decent free engines? Kind of had my hopes pinned on Gamemaker, no idea why I thought it was free.

Unity

Look into the void

Unity and Unreal are both "free", but if you make enough money you'll have to pay royalty.

Godot is FOSS but I'm not sure if I'd recommend it at the moment.

Got tired of enginedev with no graphics, so I'm battling the ghost of Ken Perlin for something basic and visible to show.

Just because you're right doesn't mean you're not crazy.

We page 13 now, and I don't see mods cutting enough threads to get us back to 12.

no u:

IMO, a bit too similar to darkness.
Keep it up though. Good to see progress.

Gravity is Earth+Dark

Yeah, I dig it. Just make sure the two icons are distinguishable from each other. Hard to tell until they're in the UI, I know.

Maybe make darkness, darker?

You can see by the design, darkness has sort of a crescent moon pointing down, the borders are darkers too, also the color is midway through dark and earth, i couldn't get any closer to earth without looking goofy

...

or i'm not doing the multithreading right
anyone know if this is how it should be?
goal of the program is to calculate the determinant of an n-sized matrix
if i have enough spare threads, i calculate the submatrices on a new thread, otherwise i calculate them on the current thread (i think)
the whole thing is being calculated recursively
public class Matrix implements Runnable { public static int threadMax = 1; public static int threadCount = 1; public double[] elements; int size; public boolean separateThread = false; public void SetElement(double val, int x, int y) { elements[size * y + x] = val; } public Matrix(int _size) { size = _size; elements = new double[size * size]; } public double Get(int x, int y) { return elements[x + y * size]; } public double Determinant(boolean quiet) { if (size == 1) {//shouldn't really happen unless someone specifically enters a 1 sized matrix return elements[0]; } else if (size == 2) { if (separateThread) { threadCount--; } return elements[0] * elements[3] - elements[1] * elements[2]; } else { double sum = 0; for (int i = 0; i < size; i++) { Matrix sub = new Matrix(size - 1); if (threadCount < threadMax) { Thread t = new Thread(sub); sub.separateThread = true; threadCount++; t.start(); } for (int x = 0; x < size; x++) { if (x != i) { for (int y = 1; y < size; y++) { sub.SetElement(Get(x, y), (x > i ? x - 1 : x), y - 1); } } } sum += Math.pow((-1), i) * sub.Determinant(quiet) * Get(i, 0); } return sum; } } @Override public void run() { // TODO Auto-generated method stub }}
in particular, this is where the new threads get made
Matrix sub = new Matrix(size - 1); if (threadCount < threadMax) { Thread t = new Thread(sub); sub.separateThread = true; threadCount++; t.start(); }

Maybe not exactly this, but you get the idea.

goblins

goblins best race

Goblins are what happen when a Jew and a fantasy nigger (orc) have a baby.

nice, 3D is fun.

so, change the angle? i would if i could, but the current design was a happy accident i had while playing with several layers of spiral gradients, i don't think i can make it look good in other angles.

Gonna write plenty about goblins in my setting, both their origin, features and powers will be quite creative
Im only keeping their appearance and name for easy recognition

idtech 3 is released under the GNU license

It shouldn't be too difficult coming from java since you run all the game code in the QVM (Quake Virtual Machine) which doesn't have memory allocation or anything.

I would recommend using Intel's extensive guides on how to take advantage of SIMD but they are all for C/C++ afaik. If you don't mind writing it in C you can call it from the JNI which will probably be faster.

the task in particular is for java

Thats why I mentioned the Java Native Interface. You can call C functions from a dll.

When you create a new thread and pass in the sub Matrix the Thread constructor calls Matrix.run()

There's nothing in your run method. So it just stops the thread.

You did create the sub Matrix though, and that means its memory is visible to both threads.

So when you call sub.Determinant(), you're calling it from within the parent thread. The child thread has already terminated, doing nothing.

What you want is a Determinator1000 class that implements Runnable and accepts a Matrix via constructor, and Determinator1000 keeps a reference to the Matrix called .boy, and has a result you'll calculate later, its .future. Determinator1000 also needs a property .sarah of type Mutex.

Then you'll have a function on Determinator1000 named .DeConnor( Matrix john ). This method takes in john then cuts him up into little pieces, creating an array:
Determinator1000 skyNet[] of more Determinator1000s and giving them each a piece of john. .DeConnor() then creates new Thread() for each element of skyNet[].

After skyNet[] is fully operational then .DeConnor() has to call the Determinator1000.JudgementDay() on each skyNet[]

.JudgementDay() needs to ensure that Determinator1000 has a valid .future before returning it. It does this via a time loop. Determinator1000 locks the .sarah to prevent unwanted mutation before checking the validity of .future to see if it's NULL (of if .future is primitive then use a separate boolean to indicates whether the .future is good or bad). Use a local hope value to record the validity If the .future, then unlock the Mutex. If there is no hope (the future is invalid) then .JudgementDay() waits on john and repeats the time loop (see Object.wait). If there is hope (the .future is valid) then .JudgementDay() can return the good .future

Meanwhile, in another Thread timeline, the
Determinator1000 run() method to be precise, it locks .sarah up, calls Matrix.Determinant(), saves the result in .future (perhaps noting the future is good), then notifies .john and returns.

To sumarize: You Chop john into little bits and feed him to skyNet[] - which is an array of Determinator1000s. Then .JudgementDay gets the result from them. You can do whatever you like with the bits that come back; It will not matter because we will have already won the war against the humans.

Shit, I got side tracked on world domination and forgot about the .boy!

Has anyone seen this .boy?

alright, i've sort of got the idea but it's not working
if i limit it to one thread, it works but slower than before
if i do more than 1 thread, i get an IllegalMonitorStateException

i made a new class that stores the reference to the matrix, calculates it's result, and when all the instances of this class are finished they notify their parent to calculate their result

the parent in this case is a matrix, if it can make a new thread it makes a new object of the previous class, otherwise it directly calculates the result of the submatrix in that thread
after it's told all the threads to start calculating it waits until they've notified it. after that it sums up their results

so the question is why do i get an error when i'm using threads

package project;public class ThreadedMatrix implements Runnable { public Matrix matrix, parent; public boolean finished = false; public double result; public int index,x; public ThreadedMatrix(Matrix _m, Matrix _p, int _i,int _x) { matrix = _m; parent = _p; index = _i; x=_x; } @Override public void run() { // TODO Auto-generated method stub result = matrix.Determinant(false); finished=true; for (int i = 0; i < parent.threads.size(); i++) { ThreadedMatrix d = (ThreadedMatrix) parent.threads.get(i); if (!d.finished) { return; } } parent.notify(); }}
package project;import java.util.Vector;public class Matrix { public static int threadMax = 1; public static int threadCount = 1; public double[] elements; int size; public boolean separateThread = false; public void SetElement(double val, int x, int y) { elements[size * y + x] = val; } public Matrix(int _size) { size = _size; elements = new double[size * size]; } public double Get(int x, int y) { return elements[x + y * size]; } public Vector threads = new Vector(); public double Determinant(boolean quiet) { if (size == 1) {// shouldn't really happen unless someone specifically // enters a 1 sized matrix return elements[0]; } else if (size == 2) { if (separateThread) { threadCount--; } return elements[0] * elements[3] - elements[1] * elements[2]; } else { double sum = 0; for (int i = 0; i < size; i++) { Matrix sub = new Matrix(size - 1); for (int x = 0; x < size; x++) { if (x != i) { for (int y = 1; y < size; y++) { sub.SetElement(Get(x, y), (x > i ? x - 1 : x), y - 1); } } } if (threadCount < threadMax) { Thread t = new Thread(new ThreadedMatrix(sub, this, threads.size(), i)); threads.add(t); sub.separateThread = true; threadCount++; t.start(); } else sum += Math.pow((-1), i) * sub.Determinant(quiet) * Get(i, 0); } if (threadCount > 1) { try { Thread.currentThread().wait(); } catch (InterruptedException e) { // TODO Auto-generated catch block for (int i = 0; i < threads.size(); i++) { ThreadedMatrix d = (ThreadedMatrix) threads.get(i); sum += Math.pow((-1), d.x) * d.result * Get(d.x, 0); } // e.printStackTrace(); } } return sum; } }}

Making your own engine costs a lot of time which is worth more than money, also it's only reasonable to make your own Engine if you're a big studio or invent new shit at Siggraph presentation level that you need to implement. As an Indie or one man army, free engines are more than enough to do the job.

Unreal has a royalty of 5% if i remember correctly, Unity however only wants a one time payment of 3k bucks or something if you earn more than 100k a year.

FRESH RBEAD