/agdg/ + /vm/ ~ Amateur Gamedev General

Spaghetti code edition

Related Places:
>>>/agdg/
>>>/vm/

Old Thread:

Other urls found in this thread:

haxeflixel.com/
twhl.info/tutorial.php?id=175&page=1
moddb.com/games/half-life/tutorials/where-is-poppy-your-first-custom-entity-part-1
moddb.com/games/half-life/tutorials/basing-one-monster-class-off-another
docs.godotengine.org/en/stable/classes/class_input.html?highlight=Input#class-input-set-custom-mouse-cursor
z80-heaven.wikidot.com/system:tutorials
z80-heaven.wikidot.com/system:appendices
z80-heaven.wikidot.com/instructions-set
godotengine.org/article/how-actually-make-your-dream-game
aigamedev.com/open/review/thief-ai/
github.com/Kazade/kazmath).
docs.unity3d.com/ScriptReference/Vector3.Dot.html
gameprogrammingpatterns.com/component.html
itch.io/jam/lowrezjam2016/topic/19276/unity-setup-request
polycount.com/discussion/126662/triangle-counts-for-assets-from-various-videogames
blender.org/bf/Autodesk_FBX_License.rtf
mega.nz/#!xdUXxb5a!BR25UspOEtSpTMygsgiw1J2BeHpPgZRywQ0N93VmOQ8
volafile.org/get/dulTPAKkpmtm/SU-100_HP_export.blend
twitter.com/NSFWRedditVideo

first for map editors

minor update
you can hit more than 1 enemy with a single punch, although usually it's 2 at best with the hook

Why is 2D development harder than it seems?
None of the current middleware engines use C++, and I'm nowhere near skilled enough to write an engine. SOS, I can use all the help I can get

non-cuboids when

haxeflixel.com/

Tried it before, didn't know haxe as well as C++, and the IDE didn't work right and bugged out.

I'm really considering using Java for my game tbh. LWJGL seems so good.

Godot 3 is looking pretty good, I'll have to try the engine again once 3.0 comes out.

what is a good C IDE for windows?
I want to be able to use new libraries and not remain limited to predefined ones. Tried eclipse but couldn't make head or tails of it.

Is it wrong to put in intentionally frustrating and patience testing sequences in to my game just to keep casualfags from playing it? If I'm being honest it's just kind of my style, I'm not really putting them in just to fuck people up, but I tend to build with high difficulty to obtain in mind when I make secrets and shit like that. Is it wrong to gate that content off knowing that only like 20% of people will ever see it (if even a number that high}?

Eclipse is garbage.
With Windows, I've used Visual Studio, Codeblocks, and Qt Creator.
Codeblocks is OK. Easy to understand. Has some basic template to get going for beginners.
Visual Studio is a mess. Setting flags, setting up include directories, libraries is a clusterfuck. The debugger is good though.
Qt Creator is by far my favorite. The project files make sense to me. Integrates well with both mingw and visual c. Works well with debugger but not as nice of views into data as Vstudio.

I also use Qt extensively, so I learned Qt Creator and it's make system as I went. Later I switched from Codeblocks and just started doing everything with Creator.

You might also look into Sublime 3. It can interface with a build system, does really good syntax highlighting, can organize projects. It's fast too.

Theres an alpha out right now.

I should have specified this earlier. I mainly want an IDE for a compiling/building. I've only used Turbo C++ when I was learning C++ in school so I am not exactly experienced with modern ones.
Not gonna touch Visual Studio because it asks me to make an account.

Good casual filters test for skill or strategy, not for patience.

Well, okay, mine test for both, but I'd say it requires patience (practice) to get good at the game.

So you mean skill? Thats fine. Unless by practice you mean trial and error to solve a puzze or something, in which case thats just bull shit.

I just finished implementing a somewhat SS2 like climbing system. Just need to tweak it a little bit here and there, refactor and commit.
Feels good when shit works as intended.

I'd say you want to design a challenging game without fucking the players over. How? I guess that would depend on the specifics of what you have in mind.

I think as long as the whole game is not based around patience, it's fine. But, there are even really good games that are based mostly on patience (Thief, Hitman 2).

What kind of game are you making?

Nah, if you're skilled enough you could probably nail it first time. I'll keep that in mind though, for now my games are pretty simple but when I get to making RPGs I'll try and make the puzzles not obvious but so the player understand what they're doing and what the possible outcomes are (and if an outcome is unexpected I won't make it negative unless there were some obvious signs it would be).


Okay, I think I've got that (more so because I built a loophole into my game that the clever will notice and can exploit for unlimited continues).


Right now just a simple block breaking game where you can win 1 of two ways, 1 is quick and easy by hitting a single special block, the other is to break all the blocks. If the player manages to break all the blocks all the levels they get the secret level (the hard part being avoiding the special block that ends the level and holding on to your limited lives and continues while breaking the larger levels).

I mostly ask though because I'm considering my design philosophy for the future.

Reposting from last bread.

Feeling comfy, in a dark kind of way. Where did you get that music? I'm assuming you made it, since it sounds like it uses the default FL Studio kick and snare sounds.
Authentic.

The more generic it is, the more A E S T H E T I C it is! Jokes aside, FL studio makes music creation ridiculously easy. I'm glad you like it, I plan on releasing a lot of the music I make as free to use after I'm done making the ost for this.
The "comfy in a dark way" thing comes from my autistic need to model the music and atmosphere partly off of old comfy as hell flash games which usually had like three short looping songs that were amazingly chill.

How does Blender compare to Maya? Installing Maya on Linux is a pain and Blender just works. I am especially interested in the rigging/animation part.

This is my personal favorite

non cuboids are easy for me to implement, i just dont have a reason to implement them in this stage. The way this works is it takes a unit model of the brush and runs it through scaling and translation matricies, so i could change the unit model to a cylinder, or anything really.

Since I've seen some people using Source in these threads, I figured I might as well ask here since googling around hasn't helped me. I recently started making maps for Half-Life 1 and Opposing Force. I'd love to make a larger scale mod but I keep running into weird little road blocks. The Valve developer wiki seems to hold a mish-mash of information for Gold Source and the newer Source engine, so its hard to tell that's actually applicable to what I'm working on with hammer 3.5.

Basically, I want to set it so that certain enemy AI groups are friendly and vice versa. I've seen this happen with other Gold Source mods, IE, friendly human soldiers, or hostile security guards. But the only way the valve developer wiki states how to do this is via the ai_relationship entity which doesn't seem to be present in Gold Source. Do I have to make a whole new NPC and set them in c++ to be friendly, or is there an easy workaround that I'm just not noticing? Honestly I'm really new to this, the most I've ever done is make a few Doom 1 maps a few years ago.

short answer, no.
games are bound to be unintentionally frustrating, specially first projects. PLUS you are already at disadvantage, most people will not know about you or your game, from those who do most will not play it and even less will finish it, at that point you might as well not exist, you could go inawoods to masturbate caterpillars and have a bigger impact on society than whatever you try to hide from casuals

I've used both. Used Maya in a school setting to complete a couple animations but I still prefer Blender for gamedev. They're both great. If you JUST want to rig and animate, blender will work great, but don't expect any of the bones or keyframe data to export over to Maya if you ever decide to switch when you get to the rigging/animation stage.

animaton and rigging tutorials for blender are fucking scarce, however animation in blender is fucking powerful, pretty much every input field can be animated, the curve editor has it's own set of modifiers and you can do physics driven animations (weights can be set by object, volume, density and even meta balls)

Its right here in monsters.cpp. You have the Half-Life SDK, and visual studio 2003, right?

Also, important links:
twhl.info/tutorial.php?id=175&page=1
moddb.com/games/half-life/tutorials/where-is-poppy-your-first-custom-entity-part-1
moddb.com/games/half-life/tutorials/basing-one-monster-class-off-another

I have the SDK but now Visual Studio 2003. I'll grab that and take a look there! Thanks for the links as well, I'll read through those.

agreed that animation in Blender is super powerful. do keep in mind though that physics driven animations made in Blender don't carry over to most game engines afaik, you'll have to handle stuff like that in Unity or Unreal or whatever you decide to use. The only workaround I found was to rig the cloth to ik bones and key each frame of the simulated cloth, but even that wasn't perfect and I could never get it to work right. I also found that it also doesn't play nice when you try to transition between two animations, should you choose to go that route.
basically just keep the cloth simulation to the engine, I wouldn't suggest trying to make cloth simulations unless you're making an actual rendered animation. I might be wrong though, if someone figured it out let me know

you sound like an impatient casual.

I haven't lurked an /agdg/ in some time and I stopped just like making game. Have any /agdg/ games been released or had a lot of work done on them since around April?

no, this is a thread for dead dreams

I've recently released a mod and a little while back did a tech demo of a platformer engine in GZD.
Currently in a "What should I make now?" phase, playin' vidya and trying to get inspiration.

What made you stop? Your game's still waiting to be made, user
I've been gone a while too but this past week I only remember seeing the fairy game and that one jet set fightan game make progress. lots of grass cutting.

Do we have a list of 8ch's agdg games anywhere? I know the other agdg does but I haven't seen any for this one

Having a list might be a good way to guilt anons into working on their shit

I've done it. It's honestly quite nice at first because things start off easy and you have much less to worry about than with C++, with the garbage collection, no pointers and all. However you quickly hit the limitations of Java. Most libraries are C++, and the bindings are always either nonexistant, outdated, or just partly functional. I hit roadblocks trying to implement AssImp and Bullet in Java, while in C++ it's all native.

Then there's also the performance difference of course, as well as some other technical limitations of Java. also, in here it'll always be >java


Visual Studio, but only after a good hard pruning of the shit features that are enabled by default and slow the thing down to a crawl.


Is that a default source feature? If not, nice touch


What do you think about 3DS Max for animation? I had it lying around and switched to it after Blender didn't export properly, does the job pretty well so far.


Maybe the list could be on the dead Holla Forums wiki

I stopped because I was shit at programming and had gotten myself into a mess of spaghetti code trying to do something too autistic for my shitty skills. I picked it up again a few days ago again and fixed all those problems though so now I'm just at a stage where I can add more features.

It's supposed to be a realistic text-based and turn-based version of The Sims set in the 1800s. It still mentions things like exact health and everything in an inventory at some points but I'm planning on changing it before it's finished to something a bit more immersive.

Pelles C

mite b kewl

I have some code that draws a crosshair on screen, pic related, but before I edited it into what it is now it was twice as long and I was repeating myself a lot. However as it is now it's slower due to all the if-then checks. Because this is code for the crosshair it's called every frame as the crosshair moves around the screen. Should I just keep it as it is now with it's 80 if-then checks per-frame or should I have it be twice as long with zero if-then checks and I'm calling different variants on the same 3 lines of code?

Do whatever is faster, and if you have different variants of the same three lines of code, extract them into a function.

I think you need to rewrite this section of code. 80 conditional statements for calculating and drawing crosshairs to a screen seems a bit excessive. I don't really understand why you need this code in a for loop; you need to update the crosshairs position once per frame and draw them, but this code does it eight times in a single frame.

working on the unreal default rig…
the left hand and right hand ik handles are parented to a "weapon" ik.
assume that its some sort of socket that i keep hearing about in the you tube videos i watch trying to make seance of how much i get to fuck with the rig.
all the videos are adds of how great unreal is to work with rather than specifics on my problems. (what do i actually need to feed unreal for my rig to say that its the default rig and not fuck everything up)
the ik setup is not bad unless you want to move the hands separately like if you only want one hand on your gun/sword/dick

about up to the ends of my wits with this fucking project.

and never before have i been more compelled to work on my own project

The crosshair is made up of 4 white lines, each with a black border around it. Each border is just a longer thicker black line drawn behind the white lines. That's eight lines total, so it loops eight times, once for each line.
I'm rewriting the code right now to be the longer and more verbose, but easier to read and faster version.

this is art. those who think it isnt are on the wrong side of history

Jesus christ, cache that to a texture or something instead of recalculating it every frame

quints

I'm designing a stealth game.
Does anyone have any tips for stuff like that or stuff that they'll like to see implemented?

I want to be able to change the crosshair size, shape, and color during the game without aliasing.


3d or 2d stealth game? What sort of setting (cyberpunk, fantasy)? My current game is a 2d stealth game but it's just me working up to making a 3d cyberpunk stealth game eventually. We could work together if you're open to it.
Aside from that, play the original Thief and build on that. It had a lot of good ideas like walking on different surfaces, low player health, keeping to the shadows (and a light gem to show player visibility), environments that make logical sense, and somewhat intelligent guards, among other things.

3D stealth game, futuristic and you play as a cyborg.
My goal is to design a game where the player and the environment balance each other out.

Write a shader for it then. What you're doing now is very unoptimized software rendering, you have a GPU to do that kind of task

Neat. What do you mean by "the player and the environment balance each-other out". Do you have an example?


Time to learn how to write shaders in Godot I guess.

The environment is meant to offer a variety of problems and the Player must have to solve them to finish his objectives. I want to make sure that the Player is well equipped to deal with those problems and that every single problem the environment offers can be countered with whatever the Player has on hands. Forgive me if it sounds very complicated, I just assumed that's how must stealth games worked in their core principle.

Many words, nothing to say

Is it possible to run live2D in penguin os?

That sounds nice and all, but this can lead to stale gameplay. If a player can solve any problem in the game with what he has on hand, many players are probably just going to take the path of least resistance and choose the fastest/easiest strategies to solve whatever problems they come across. You'll need to incentivize players to use more creative solutions to combat this.

I did think of this so I'm going to give each situation a set number of solutions that can be used to bypass them, the Player would be forced to choose any of these provided solutions. I'm also not going to tell them what solutions to use so when they figure one of them out they'll think it's out of their own creativity, this adds a level of false depth to the game.

man love
fuckin retard

so like a zelda game where you get a grappling hook to cross a gap that only exists to force you to find the grappling hook to cross?

welcome to metroidvanias.

well my question still stands what for example is going to be an obstical and what is going to be the solution the player "should" use?

Yes but you'll always have the grappling hook, you're just not told to use it here

The simplest one I can think of now involves luring a guard from a door so that you may pass it.

Just use this:
docs.godotengine.org/en/stable/classes/class_input.html?highlight=Input#class-input-set-custom-mouse-cursor

live2d != love2d

Added all the door and window frames and some railings. I made a a blueprint that randomizes the materials used for the top and bottom walls of each building. Helped out a lot. Going to add dirt on walls and more color variations for doors and windows.

so you throw a rock?

I mean fam, do you have source code? Does it have penguin binaries? Are you fucking retarded?

Yeah, basic stealth stuff like I said

Hey anons, I haven't been to an /agdg/ in ages. What ended up happening with LoliSim dev?

Also does anyone here use Godot? I've been meaning to get into it for a while now, and just now actually started trying it out and I'm liking it a lot so far.

not using a sonic gun to fire a localized echo

I'm currently using Godot.

After a week of retarded fumbling, I was able to get the basics of an game entity system in place. At some point I should really try and draw out some sort of architecture or have one central design goal.

P.S. FUCK vector iterators I might have autism but this crosses the line


Older versions of visual studio seem to be alright. Alternatively, you could skip the IDE for a text editor, and a compiler with some batch macros.

There's an /agdg/ thread up for SimLoli, he's still alive and kicking, but the thread moves slow.

Im starting to like UE4's blueprints a lot. Although I feel like it has a high learning curve because of all the scattered documentation of bugs.

How the fug do I do this? I want this block panel to align to the other face as exactly on the first one, but Blender always shifts the block thing more to the other side instead of aligning in a way that it is the same as it is on the first block. The first block is placed in the middle but this is not the case for the second block.

Blueprints? I know shit-all about UE4. What's the tl;dr on this new feature?

Hasn't it been like one of the primary features of UE forever? It's their click'n'drag programming thing.

Use instanced duplicates. Make 1 side flat perfectly. Go to Object Mode, Select that perfect panel side. Move your cursor to the very center of your pentagon? Hexagon? Press Alt+D. Press Enter. Then Press R to rotate and Z for along axis and type in (polygon sides/360). Like vid related. Good thing about it too is that editing one side will edit the others since its an instanced duplicate as opposed to just a duplicate.

Ahh alright thanks that solved it, but I had to type in (360/polygon sides) which in this case is a octagon/8-gon. The tip with the instanced duplicates sounds useful now I can actual edit the blocks even further I didn't knew of that previously. If I typed in (8/360) for the rotation it just gave me awfully small rotation value like 0.000022 or so.

Why don't you draw the lines with Polygon2D nodes and just parent them to a Node2D that follows the mouse position?

That was me, sorry. I'm the codev trying to wrap my head around Source engine by just trying to read through the code.
I kind of have a question as an amateur programmer. I have little game-making experience and I know Source isn't really the shining example of engines, but can anyone explain or show me to any explanation of how engines like Source are able to simulate and display so many fully 3D, independent objects at once? My only real experience with making games was fiddling with turn-based top down stuff, and I can easily picture something like that inside a game loop of input->process->display.
I don't really know exactly what I'm asking, but I'm looking for a bit of context to help me understand how people just like make game engine, so that at the very least I can modify this one with some conscious understanding.

...

Welp I am almost done with my Ore Mine, the only concern I have with that is getting the conveyor belt animation done, hmmm I wonder how hard it is to do it via textures which would be for me a bit simpler to do then having to fiddle with the curves stuff again. 4th image is my old model of it.

Also when I am baking meshes and I have several duplicate meshes that I don't want that it takes even more texture spaces for the same object, whats the best way to deal with it? I have like 3 identical boxes model right now and I want that only one of them uses up the texture space. And when I am splitting my objects for baking so that the shadow doesn't merge with each others should I keep like another copy of those 2 models? So that I have 2 models (low/high polygon) that it is in exploded view and another one that it is not in exploded view.

Bretty gud work on that self perpetuating nogame loop. Fuck that and start a new project or rush the current one even if it'll end up being a mistake ridden pile of shit.


Those looks very appealing. What game are you making?

I want to make a master system game as one of my projects. Anyone have an good tutorial on z80 assembly?

Fuck yeah curry

This has been sitting in my bookmarks forever, see what you think.
z80-heaven.wikidot.com/system:tutorials
z80-heaven.wikidot.com/system:appendices
z80-heaven.wikidot.com/instructions-set

inb4 pajeet fairies

...

Ok, so there are a bunch of types of objects in source, there is, the game world, which ofc is using a BSP-PVS like in quake, which means that it is stored in such a way that most of the level can be discarded so you only draw the viewable part, then, there are the entites, which are just all updated in an array or something, most of the time its just something like "can I see the player? / do I have a task?" The first one can be solved easily with a simple line-of-sight test, this is easy, just one line of sight test per entity is pretty fast when you have a BSP to do line of sight testing. Second one is easy too, you just check on a struct feild to see if its set or something. Then it has the physics objects, which is the part where valve bought a commercial license for the havok (tm) physics (tm) proprietary library that does the cool physics, normal entities just use the old style quake physics though. And then of course there is stuff like particles which im sure are put in some kind of variable array or whatever.

So, thats the run down on how something like source works, its a lot more complicated but that's basically the gist of it. Why not go look up some developer talks about how engines work? Here, watch this, its an in depth explanation of the quake 3 rendering engine, which should give you some kind of grasp as to how these things work, its embed related.

Make a shittily coded engine anyways.

you can stay honest if you tell everyone that you got bored and started a new one :^)

(checked)
Thank you, I appreciate it. I'll watch that video shortly.
I don't really need to have a full understanding to do what we need, but I think I need to have more knowledge as a programmer.

It's for a waifu sim. I just finished staircases and some building props. The buildings are randomized with blueprints so they all have different colors, prop sizes, etc.

jesus fuck that's loud as shit even with my speakers set to low u fucker it startled me
looking good otherwise

Explains the overly detailed church. Gotta make marrying the waifu visually pleasing

...

Thank you

I'm not watching an (1)'s unsolicited video without some context first

It's a parody of overly-ambitious crowdfunded games.

Amazing. I might just use this instead of Unity. I'm tired of waiting Front Mission 6 to happen.

I'll make my own turn based customizable mech game. With lewds and card mini games.

Does anyone know where 3D artists go to post portfolios and shit like that? I've been putting off starting a game and I'll have one less reason to talk myself out of it if I know of a place (and a price) where I can go to get some models.

I keep telling myself, "Sure, you can make simple models for testing purposes, but you can't make good-looking models and don't know where to get them so there's no point actually doing anything right now."

If you're just starting try blendswap or OpenGameArt first so you don't end up blowing your money hiring a modeler to make models for you if you don't finish your game. Plenty of decent models between the two, just search.

What kind of models do you need, anyway?

if $ fun = "false" and $ motivation = "false":
Me: "I've simply got to kill myself, its the only answer."

*robusts you* back the fuck off!?!

UE4 Blueprints are buggy as fuck. But when they work, they work very well.

How are they buggy AF?

Crashes on compilation, AddChildComponent doesn't work in Macros, EventDispatchers work in very specific scenarios, having to add delays to counter "which module came first" chicken-egg problem, etc.

Welp I figured out the baking procedure for such complex meshes now, I just need to group those models in their own and separate them together and make a exploded view in order to prevent bleeded shadows which looks bad when it is animated and then when I have duplicates of the same model I just don't create a low polygon model out of it, then I just need to do the usually baking procedure and it should work.

I did a test with 3 "high" polygon crates and only 1 low polygon crate and it worked as excepted.

I'm gonna do it
I'm gonna make porn

I'm proud of you, user.

As much as I like Unreal:

SLATE. FUCKING. SUCKS

Holy shit. Programming a UI in C++ is every bit as fun as you would expect. Before anyone asks why I don't just use UMG (the sane system with the visual UI designer): I'm trying to program a loading screen without any of the ugly workarounds that people use. UE4 has a system for loading screens in place, but the catch is that you have to use Slate, because UMG relies on some systems that aren't available before a level is loaded.

New enemy, the genetic horror. It's hard to focus on vacation

Why are you all not posting shit, at this rate i will have to work on my own game

I can't dev at work.

I've made basically no progress today, despite devving all day. I wasn't kidding when I said that Unreal 4 Slate is awful.

I just released an update, I'm taking a fucking break.

do it faggot.

Why is making video games so hard and time consuming? Why cant it be easier and better?

Just ask here, plenty of somewhat talented 3DCG artists. Some might even do it for free.

For commissions of fully rigged low poly models (lowpoly as in runtime usable) I usually ask $150. Animations on top of that generally go between $15-$50 depending on their complexity.

You can always be an ideasguy for zero effort required

oops

There's a lot of artists, and their portfolios on artstation.
They range from amateurs to pros.

What a defeatist mindset.
Half the fun of developing games is learning, iterating on lessons learned, and putting this together into a tangible result.
3D modeling, sculpting, and animation are all pretty entertaining, and I consider them the relaxing alternatives to programming.

Interesting read on How To Just Like Make Game from the lead programmer of Godot
godotengine.org/article/how-actually-make-your-dream-game

Wow, the advice that the author gives in that article is pretty shocking. I've never really understood the thought process behind "code things fast and dirty", but thinking about it from a game design perspective actually makes sense. I've left countless projects unfinished simply because I'd lose momentum and get autistically caught up in things like making premature optimizations or doing things as efficiently as possible. As someone who has a degree in CS, quickly prototyping a game with pajeet code is going to be a bitch and go against all the shit I've learned. Fuck.

that also hit me hard, It made me realize i stopped making progress the moment i tried to do things "the right way"
also i going tomorrow to see if they do incentives for games here and the requirements to apply

In UE4, I have a blueprint actor that is using SpawnActor to spawn some other Actors.
No matter what transform I supply, it always places them at 0,0,0.
What the fuck is going on?

Did you check "use manual transform"?

Interesting. The bit about coding quickly rather than "correctly" is at least mostly true, from what I've encountered in my admittedly limited experience. I still don't know where to draw the line on optimizations and cleanliness, though. I need to be able to read my own code six months down the line, and I can never know if my retarded implementations aren't going to work with something else I make later, since I don't know how some things work. For instance, how can I justify adding all of the other features that I know how to implement, when I haven't even learned how to implement predictive movement yet? Surely that must come first, right? Or am I over-thinking it? The grind of trying to preemptively learn everything that I'll need to know before I need it is getting to me, since it slows down being able to work quick and dirty on things that are actually fun, and thereby see progress.

No, but I think I tried it with that on once to no avail. I'll try it again.

Octrees are fucking hard, at least dynamically updating stuff within them and including boundary cases.

this is my progress on the map editor, as you can see I have implemented camera controls, as well as 2d scaling and panning.

Easiest way is to have overlapping cells then there's no edge cases.

That didn't fix it. The problem is that the spawned actor is being created with a transform position of 0,0,0 despite the transform position going into the spawnactor function very clearly being non zero. This is before even attaching any meshes to it.

What's the basis for designing AI in stealth games?

I recommend aigamedev.com/open/review/thief-ai/ as a high-level overview of Thief's AI.

Because teams of multiple people have been required to get a game finished in a reasonable amount of time since the NES days.

i wish i had a reaction image for this but they also were on the HD that got nuked

BACK YOUR SHIT UP user

If it was really work you were happy with, you would have backed it up

...

...

Working on optimized code a lot is not entirely a problem either, as you become faster at writing optimized code. Quick and dirty pajeetcoding is mainly for
prototyping new functionality, which we already did to such a degree in the team I last worked in, that the designers were prototyping important functionality
themselves with blueprints.
I still have nightmares about some of their implementations, but the project wound up pretty decent after a big refactoring.
For the project itself, you don't want to be full-on pajeetmode though. Pajeetery is a spectrum ranging from ass-pie heh to full on street shitting, and with
experience you can feel where on that spectrum a piece of code should be. Some of the code in a vertical slice should try to be really good code, code in alpha can
be mid-tier shitty, beta you want pretty good code, and while working on gold you go all over the scale as you implement fixes and small features that range from "I
should probably get this working" to "oh god this bug gotta rewrite this entire feature fuck".

The part about good interaction APIs is also very, very important. It doesn't matter how shit your code is, you have to encapsulate it so other systems can work with it and that encapsulation code should be clean as that's what you're most likely to use/read in a few months time. And remember to comment.


If this is only for that particular actor, it's most likely a problem with your beginplay or construction scripts on said actor. If just spawning a cube still places it at 0,0,0 then please show your spawning code.


GIT GUD YOU IDIOT

...

i told u to stop and pick existing solution

And he's not going to listen, because he does what he enjoys doing. Do you even know who he is, or his previous project?

pretty cool. I take it you are doing all of the 3D math yourself, and not using something like GLM?

How are you storing the 3D primitives for the map or scene? Are you using BSP, are you discarding geometry that won't be visible?

...

Good work, user.

Wew, that took long enough.

I'm not sure I know how to climb on top of obstacles, can you tell me?

Are those Brackeys videos that everyone recommends in these threads any good? How soon after completing their C# tutorial would I be able to start making vidya? I like knowing how long these things should take.

I see people unironically asking which language to program muh_gayme in. Since Java and C# are both C++ applications, why not just cut out the middleman and program directly against the base runtime of C++?

After all the fastest code is the code that isn't there. Removing the intervening runtime between your game and the hardware means it will execute quicker.

There's this thing called "compilation."

Ofc, but that still doesn't obviate the intervening runtime. Besides precompiled is best compiled.

Don't bother with his C# tutorial for beginners, it's boring nerd stuff with compiler and console. Instead go for Unity tutorials for beginners where he talks about both programming and use of the game engine.

Depends on your current knowledge of programming and intelligence. Only you know answer to this question based on how much you understand from what you're watching.

I did the first 5 videos in his normal C# series and only needed to rewatch once to grasp what was I learning.

That runtime is there no matter the language used, and the overhead from actually drawing shit is probably going to be several orders of magnitude more taxing than any of your game's logic.
Unless you've got some really shitty logic.
Or you're trying to enginedev, god forbid.

C++ is only really faster than, say, C# when you're intentionally using its memory access shit and the like which C# forbids.

Make the ledges thinner and pull the middle support beams on the ceiling a little down so the base is straight.

Maybe, but there's a pretty marked performance difference between executing bytecode in it's runtime, and machine-specific binary code straight in the hardware. Good point about the graphics code. OpenGL ftw.

I think I'm gonna make my first game basically I-Ninja's ninja ball mini-game if it got a spinoff game dedicated entirely to that

i'm going to do both in the compiled maps. But the 3d data that the editor uses wont have any culling or BSP or whatever. I am doing the 3d math myself, no GLM used because GLM is in C++ and I am writing this in C. I actually used openGL as a math library somewhere in the code:
//use openGL as a math library ;-) glLoadIdentity(); glRotatef(g->theta_x + 90.0f,0,1,0); mat4_t mat; glGetFloatv(GL_MODELVIEW_MATRIX, mat.m); vec3_t vel = {0,0,1}; apply_mat4_to_vec3(&vel,&mat); g->camera_x -= CAMERA_VEL * (float)t->tstep * vel.x; g->camera_y -= CAMERA_VEL * (float)t->tstep * vel.y; g->camera_z -= CAMERA_VEL * (float)t->tstep * vel.z;

...

So what have you been working on for 6 months?

A 3D platformer, acrobatic, skill oriented, tight gameplay and obstacles, but with a dream-like world to explore. Kinda like cloudbuilt + alice madness returns.
Tempted but wouldn't risk showing it since some of my co-workers are normalfags.

Please help me.
I've been trying to create a behavior table for my stealth AI for a while now and I have no basis to follow. I read the ai thief review but it seems to only cover the technical details. I don't even know where to start making a basic stealth security system for a Player to face.
How do guards work in real life?
How do companies protect their assets?
How does everything work together to prevent intrusion and theft of data?

Bad idea. There's a reason guards in games have pathetically short vision and are as forgetful as retarded goldfish. Having guards who go everywhere in pairs, watch each other's backs, radio in immediately and wait for backup, etc., is not fun. I assume you want your game to be fun. Don't mimic real life too much.

Anything beyond that and I'm going to have to ask a bit more about what style of game you're going for.

...

That's disgusting. Just use CLAPACK or ATLAS or something (GNU Scientific Library is best if you are using GPL, otherwise you won't be able to use it). If you want something more modern and possibly better-maintained, kazmath is apparently really good and it's pure C, but I've never used it (github.com/Kazade/kazmath).

Most importantly, those OpenGL functions are deprecated and you could run into issues using them alongside modern OpenGL. They're also very likely to be unmaintained and slow in modern drivers. I wouldn't suggest that anybody use the deprecated OpenGL functionality for any reason other than maintaining a project written with OpenGL 2 or below.

If I were you, I'd use GLM and C++ to write just the subsections of the code that need linear algebra. You wouldn't have to switch to C++ entirely, and it would make your life a lot easier.

gynoid user?

No.

Please, calm down for a moment and think about why I am doing this, I think you're having a huge knee jerk reaction. Let me explain the project:

I am using OpenGL 1.1 to do all of the graphics in the map editor. This is because I can write OpenGL 1.1 code really fast. I'm not ignorant about modern graphics programming, or optimization, the main game uses Vulkan. This editor isn't using modern OpenGL. You could say that this is slow, and I know that it's slow. The only computer that it needs to run on is my computer, so, I don't care about optimizing. It runs a simplified view of the level with no culling and no effects or anything, on my overpowered development hardware. So, I just want to program the editor faster.

Now, you might think, that its dumb if I was using legacy OpenGL as a math library, but this is deliberate. Its not like I don't have my own linear algebra routines. (on the actual engine, I dont even use the OpenGL built in matix operations, I only use glMultMatrix() in the OpenGL renderer) The reason I would want to grab the matrix out of openGL is this: The way I implement the orientation of the camera is like this: glMatrixMode(GL_MODELVIEW); glLoadIdentity(); glRotatef(g->theta_y,1,0,0); glRotatef(g->theta_x,0,1,0);
Now, lets say that, I want to move exactly forwards. The way you do this is you apply the vector (0,0,1) to the model-view matrix generated by those calls. Now, I have two options:

First music thingy I've done in a while.

That would make it so much more fun, I don't know what you're talking about. So many stealth games these days make it so that the easiest way to sneak through a level is by picking off guards one by one.
Having guards who work in pairs and actually have some semblance of brains would make gameplay where "ghost" runs are preferable and violent altercations less desirable. Instead of this stupid fucking trend where everything just feels like a slo-motion serial killer simulator. We need more stealth games where if you're sent to assassinate a guy, the easiest path isn't to silently kill an entire military base.

me on the right

I don't know shit about music composition, but it feels like you're missing a high 5th note to come out right after the 4th.

Are there any good resources for learning Godotscript? Last time I tried doing anything in Godot I just got confused, but since the free version Game Maker Studio 1 is no longer available and all I've got is a pirated copy of GM8, I think it's finally time to buckle the fuck down and learn to code the hue way.

why

Look for resources to learn Python, not godotscript. I used to know a good one but it went behind a paywall.

Oh okay, nevermind then. I actually know C#.

It just hit Alpha like a couple of weeks ago, and I'm pretty sure they said release is still a ways off.

But I agree with in that people should just learn Python until then since it's so damn similar.

How would I make turn based combat in GameMaker?

I haven't the faintest idea how I would even begin. The problem is keeping track of turns and looping it until the combat ends either because one side dies or the player runs away.

...

Very vague question, I know, but does anybody know that function in unity that compares two Vectors and returns 1, 0 or -1 depending if they are perpendicular or opposite?

I cannot fucking find it or remember it

found it
also check em

...

(checked)
nice fuckin memerino

I mean it's a single function and the answer to a question posted in a thread that will be dead in 24 hours and nobody will ever see again, does it matter?
It's this docs.unity3d.com/ScriptReference/Vector3.Dot.html

What if someone decides to screencap your question to ensure it will be passed on into the future?

where can I learn how to make vidya ? what should I learn ?

welp


Download unity and dick about

There's many ways to make vidya. Either you learn to program and then make games with those skills, or you download one of the many engines and then google tutorials until you get it (and learn some programming along the way).

For example, Cities Skylines was made in Unity3D engine, while Minecraft was made with just code.

we have an archive, and believe it or not we are on a popular website


make pong
then tetris
then a simple rpg

Or make something interesting so that you don't lose all inspiration to make a game after 2 hours

you could also follow tutorials untill you feel comfortable

use the godot meme
just learn programming and then focus on only making games

Pick a game engine and watch some tutorials. When you'll start to understand how to do something on your own then dick around and try if your ideas work. If something doesn't work look for solutions. Treat gamedev as fun game itself and don't try to impress anyone but yourself and you'll make it.

I'm currently rotating my character by setting transform.rotation, which takes quaternions to rotate in the direction of movement.

BUT I'm trying to rotate the player to face directly a wall in case he's walking into one.
How do I transform an Angle value into a quaternion to feed it into the transform.rotation?

Using unity

apparently it's just as simple as to setting lookrotation to the negative of the normal of the raycast hitting the wall
huh

I am probably gonna give unity a try

Also in my personal experience tutorials aren't that useful because you end up just copying what they do. I think the best way to learn is make your own character controller by piecing together info from the docs and google searches.

Start off with basic movement, learn jumps, understand how to check for ground and shit like that, you'll eventually absorb enough knowledge to do anything you need.

Unfortunately once that's done you've gotta learn about data and shit which is very boring.

...

Laaaaame

bridge interior not yet implemented
instead have an oh&s hazard

That's pretty great, are you using your own character controller?
Is this even unity?

I'm not entirely sure what a character controller is I know that unity has some type of component called that, but I'm not using it. It's just a bunch of trig and ravioli code with assistance from the rigidbody.

Welp at least I saved 200 yuropoor bucks that way, is Blender even using the Graphic Card at all when editing a model? I even tried searching for a few solutions on a search engine and one of them mentioned is by using bumblebee and CUDA directly on it but bumblebee failed to work on me since according to this program I don't have any integrated GPU at all and using the CUDA command line didn't made any difference.

I call character controller anything used to control the character, even if it's written by you, not sure if that's right or not
Seems to work pretty well

Tell me what seems better for a platformer, I'm not sure.

or

If you're using Cycles, you have to turn GPU rendering on.
If you're using Blender Internal renderer, it's always done on the CPU.
General modeling is done on GPU, but they're using the old OpenGL 2.1 Direct Mode API to maximize backwards compatibility.
This will be changing in Blender 2.8, where they will be utilizing more modern OpenGL features to get better performance (and a lot of cool new features!)

In what way, head crash? computer burned? you may still be able to recover it, it just gets expensive.

When I lost my shit messages like this gave me false hope, don't do that to him

Depends on the platformer.

Make sure the character doesn't grab the ledge if the jump can bring you on top of the platform.

I think the first option is better because it gives more control to the player. The second options adds some waiting time in certain situations and it might be annoying.

If the player wants to specifically grab the ledge and then jump again or something he won't have to wait additional time it takes to start falling down.
If he is not interested in what the ledge has to offer and just wants to get to higher platform he can just jump the second time onto the platform.

Or you could make it that if you hold down the jump button, the character won't try to grab the ledges so the player decides which is more suitable for a given situation.

False hope? like I said it depends on what happened, I've recommended people send HDDs to a place called datalab, it's expensive, or at least it was last time I recommended someone use them, but they can generally recover data from a dead HDD, 1500 hrs is a lot of development time, surely it's worth at least $1 for every hour, it's cheaper than that though.

It's Mario64-ish.
Also I have no idea how to check if the jump would be able to get him to the platform.


Yeah that's what I was thinking too, probably gonna go fior the first one

Me either, I'm just pointing out that it would get very annoying if you try to jump on a platform and the character decides to snap on the edge to grab it instead.

Just finished first implementation of a primitive entity component system using gameprogrammingpatterns.com/component.html
as an example. Now to play with storing the gameObjects in vectors and iterating over them, and my final test, to make movable objects. If I can do that I can finally rest at ease for a while.

I'm very close to ditching C++ entirely and writing this shit in C.

Anyone knows how to get colour IDs from Houdini to Substance Painter? I know how to get different texture sets, but I want the colour ID map as making a separate texture for each small section would be performance suicide. Painter has 4 ID methods: Vertex Colour, Material Colour, Mesh ID, and Polygroup/Submesh.

Vertex Colour does not work because I'm not baking from a super high res sculpted mesh, as such I get a gradient between different colours which isn't what I want, I want a clearcut map. Material Colour doesn't work because if I assign a different material of different colour to each part, then Painter will interpret it as another texture set, because that's how you do texture sets. Mesh ID is obviously not what I want as that is just the texture sets. Polygroup/Submesh sounds ideal but I can't find anything about it in Houdini, and it seems just simple groups aren't cutting it (tried it, no dice).

I guess the tl:dr is: How do I make Polygroups/Submeshes for Painter in Houdini?

Well that is weird, if general modelling is done on the GPU then a GTX 970 would have performed still bit better like that it is not so laggy at all, right now I have a GTX 570. Still at least I got the opportunity to test a newer GPU before forking out 200 yuropoor bucks just later to find out it did not helped me.

I hope that also means edit mode can profit from this change.

I looked at their code news section and I haven't found anything interesting at all, mainly I am interested in improved editing, UV-tools and anything else that is needed for game/mod development. But so far they only reported news on dependency graph, PBR rendering, and Grease Pencil. Oh yeah they made a video about some new UV-mapping "algorithm" but for some reason they tested it on a organic mesh only, they did not made any test on hard surface modelling with that new UV mapping thing. In other words I am not really impressed on the current 2.8 development but apparently its still in Alpha stage or whatever so I guess time will tell if the 2.8 version will be worth it or not.

Well its nice that they have finally a GPU based rendering even if its sadly limited to Cycles only, but I don't really use rendering much apart for doing rendered sprites and maybe once in a blue moon a smallish rendered scene or so. Most of the time I use Blender for making models, I don't use it for painting since the lack for fixed brush scale triggers my autism.

That falling is pretty god damn funny, not gonna lie.


Just make it so the character grabs onto the ledge if the player is affecting the character's movement direction towards the ledge, problem solved that's what I did

Well, now it just werks.™ I can spawn shit wherever.
I guess UE4 just doesn't like it when you try to spawn a plain Actor.

checked
I supose you are not refering to Holla Forums's ID.

my toaster was dead for 4 days because of a shitty harddrive, and i've got exams next week
there's no time to dev like never

The latter seems like it could get annoying fast.
Imagine you are jumping up a series of blocks. If you jump at the edges, you'll grab the ledges unnecessarily every time.

i recommend this guy instead, way better

I'm thinking unity would be good for it. kinda looking up C# tutorials and everything

That prat fall is going to be what makes your game

r u me?

What's with the sudden influx of people that want to start making a game?

i hate my family and want to defoo but im a fucked up neet with social anxiety

gotta make a game and be succesful with it

Have you seen the amount of trash coming out this holiday season? Probably a big part of it.
The red tick-marks are for games that might not be shit

GOTY guaranteed dude.

Knack 2 baybee

I appreciate your position fully, as I've been there, but gamedev isn't going to solve that problem effectively in the short term, which is what you need. Learn a trade, or anything else you can start making money with in the very near future. Also, social anxiety is something you have to learn to overcome, because it's never going away. Vid related.

So they aren't even bothering to remake them now? Just straight up rerelease and charge people again?
Boy I sure love video games

That's probably Virtual Console aka pay 4 roms.

I don't know about the rest of you anons, but having grown up on these boards I'm now about to hit 30, it's fucking time to get to work and stop shitposting all day to make the games I've had in my head a reality. I'd imagine I'm not the only one hitting this point in their lives, so that's how I'd explain the sudden influx. That and the absolute shitfest vidya is right now.

General state of the industry as well as summer break, would be my guess. Most people sit idle and want to keep themselves busy. I don't think it's bad, I just question how many of them are going to stick around.


If the guy truly wants to make a game, it might help him long term though. But you're right, he shouldn't bet all of his horses on that one thing.

That explains the recent crackdowns. We will always have a choice though, if they like it or not.

I'm by no means saying that he shouldn't make a game despite being in a tumultuous position. If I did, I'd be a hypocrite. I'm saying that making a game isn't an effective strategy for getting out of that position. However, it might give him something that makes the suffering involved with escaping worth it.

No, but learning WebDev alongside GameDev might. It takes about the same amount of effort as learning to program and use Unity properly. At least then he'd have something better than Retail.

You could learn both Javascript and C# to use for both situations.

...

Agreed. It should be part of his goal to improve himself, but not the only one.

I find it kinda fascinating how JBP is resonating with so many people lately. There's something obviously wrong with our generation.

Probably man
Does this sound like you?
If so we're probably inter-timeline drifting.
stain or stein?

search for dot product or something, I don't use unity

So after an extended break, managed to come back and get the jiggle to work correctly on the full rig. Now need to do the hair…

She looks like she has a cold.

Getting gud, mate. Update at /3d/ once in a while.

Does anyone have any good tutorials for LMMS or should I just try running Mixcraft through wine? I'm trying to move away from working with midi files because the midi synths interfere with pulseaudio and it's a nuisance to deal with.


stein

Close, maybe distant cousins?

Been making a Mario clone with SFML, but I'm hardly invested in it. I started it just to prove to myself I could do something in C++, but I haven't really enjoyed programming in C++. I keep eying C like she's that MILF that jogs down the street sometimes.

yep
yep also
FUCK, do I feel you there. From here on out I'm gonna try and use "orthodox C++", basically what that is, is any C++ code that can be compiled with a C compiler, or is C++ without OOP.
Definitely not me there. I'm not an out there ideafag, but I have an idea that's more than do-able in 2D, if I wasn't a retard.

SFML has C bindings, takes some getting used to but it's what I'm working with right now.

Well best of luck and keep hanging in there. I'm sure you'll turn the idea into tangible form eventually.
I think I'm going to make the actual jump to C. After checking the job listings in my local area, not a single fucking place is hiring anyone with wanted C++ knowledge anyway, so there's no point in learning it for the chance at a career with it(like I was originally intending). I can probably extend what knowledge I've learned of C++ from this project anyway should I ever get hired somewhere, but I'm sure I'll piss of my team for not knowing Modern C++.


How would you compare it to SDL2 or Allegro? I was thinking of switching over to a pure C library when jumping to C.

Probably so. Modern C++ is very frankensteined together, with lots of use of auto and std:: things.
If you want to use languages currently in demand, C# is the most dominant, and elixir has been becoming more popular for company projects over the past year and a half thanks to it being purely functional and concurrent.
Why not SDL at that point?
Good luck with your jump btw

SDL2/SFML/Allegro are all pretty much the same shit. They have the same shit coding in a lot of parts as well. But at aggy daggy levels this shouldn't matter.

Never used Allegro, and I've only got minor experience with SDL2. First difference to me between SFML and SDL is that SFML is a lot more readable and has less boilerplate than SDL. The main reason I went with SFML is because it doesn't require setting up another separate library just to have audio (I'm pretty sure with SDL you needed to install something like SDL_Mixer or whatever, maybe they've changed that since then).

How do i get my walking animations to not look like sliding on ice?

For reference

they look alright to me, maybe make the hips move a bit more
at this point you're fine, if you want to improve it further look into foot IK

layered animation and IK legs

nigger are you blind


How's your character moving?

It's not the animations, well maybe a little bit, it's how the movement is handled.
The character would be sliding either way even if it had good animations

Posted the second webm for reference because they don't use any of the new school animation blending techniques like in AAAs these days, no layered animations and nothing except for basic IK, yet it doesn't look bad at all.


Walk cycles without root motion, a 2D blendspace with a walk and a sprint animation, default third person UE4 character controller.

It can many things. It can be getting the walk speed to match the animation. If someone looks like they're doing a normal jog but they're going super fast it won't look right.
Dark Souls 3 also uses IK(as well as most games in the AAA space these days) for characters so there's that too.
But for your prototype I can point out some things that need to be changed. Like the player should lose speed when landing after a jump(this is the biggest complaint I have really), and you should lose a hint of speed when turning(Like barely noticeable). Might even help to give them a slight lean when they turn because that's how it works in real life and many games have this.

I've fucked around with SDL before and you're correct. You gotta bring in and initialize SDL_mixer before you can use audio. I guess maybe I'll give CSFML a shot then, I really liked being able to load in different image filetypes without doing much extra to make it work.

I was just thinking of your project earlier today. And thought you probably dropped it. You have to exaggerate walk animation to some degree or it won't look right. There's force involved every time you kick off the ground in a step that shifts the entire body around. In particular, the rotation of the hip and the movement of the upper body - to some degree, the upper body has to bob up and down as you move. Go look at the Animator's Survival Kit over in /loomis/, that book's really good when it comes to walk cycles.

the walk speed doesnt match the animation

also dark souls animates run cycles where the feet are angled a bit outwards so running naturally eases into either direction without the need for layered animation

speed up animation cycle, or slow down player movement. Try the latter first, speeding up an animation cycle can make it look like utter dogshit

I'd suggest SDL since it has more tutorials, and IMO their documentation is pretty solid.
Most of lazyfoo's tutorials also applies pretty nicely.

Your reference webm in is a good example, yeah. Take a look at how much the character's upper body swings around (and the braking animation is also a nice touch). Also, next time, record the character running in a straight line instead of spazzing around like a retard, it makes it way too hard to see how your animation actually looks like.

sorry if this turns into a doublepost

speed up animation cycle, or slow down player movement. Try the latter first, speeding up an animation cycle can make it look like utter dogshit

I'd suggest SDL since it has more tutorials, and IMO their documentation is pretty solid.
Most of lazyfoo's tutorials also applies pretty nicely.

The default controller is the problem.
Go check how that's handling animation, if it's moving with physics or whatever else.


Look at the jumps, the character literally slides to the side, it's not the animation.

Go check how that's handling movement*

I guess you meant to reply to me. Didn't drop it, just had to spend a lot of time on extracting my shitty classes into proper, separate components, communicating with delegates. I also extracted the hit detection, lock on and so on from blueprints, it's all C++ now, that was a surprising amount of work. Enemies can equip weapons now, and attack the player. Pretty much everything else was invisible under the hood improvements. Thanks for thinking of my project, here's a progress webm fresh off the press

When im walking sideways, capturing it and going through frame-by-frame, it looks synced to me, comparing with the tiled floor, on both the run and the walk. I think it's the turning specifically that causes it to look like shit, and it makes me think it's something fundamental. The animations are the default UE4 ones, so they're at least serviceable


That might be it, yeah. What are the options im looking for? It's all in one fuckhuge .cpp file.

Hard to say, the only thing you can do is try to track it down by starting from how the game handle movement inputs and seeing where that brings you

Looks great, can't wait for the inevitable VaatisVidya video about this.

Went to look how Mario sunshine does the grabbing ledges stuff, and like an user said, Mario only grabs on if the player is tilting the analog towards the wall ONLY when mario's current height is lower than his older height.

So I guess that's also how I'm going to do it

or simply velocity.y < 0

Oh fuck that's much smarter than what I was about to do
thanks

Make the switch, you wont regret it. C is great.

I have concluded that the builtin UE character controller is trash and looks bad even at stock settings in the 3rd person template, but i've got no clue how to improve on this or even what's the issue. It's something with the turning, maybe the midair turning or the high rotation speed.

Just like that without fancy parallel universes?

Is not actually stupid what you were thinking, just probably unnecessary in this case. Checking the current position against the last one is a good way to find out if a kinematic character is actually moving when their velocity is a non-zero vector. In kinematic movement (non-physics engine driven) velocity is effectively the current as yet unapplied force where current_pos - last_pos gives you the actual applied distance of the last tick/frame. This can be an important distinction, for example if you are pushing against a wall your object's velocity might increase but your actual movement will remain unchanged. So for checking if a walking animation should play, you should be checking the actual distance changed rather than the "as yet unapplied velocity force". Otherwise your character might run against a wall without going anywhere when they ought to be just standing still.

Fuck this Chinese POS, go with Unity.

it's definitely the turning
make it turn overtime, not nearly immediately
or maybe make it depend on how far he's gone forward

your toaster can't run unreal, can it? don't worry, there's no shame, we're all there

You should probably make your own controller, you'll learn a bunch
I personally think this could be fixed by making the object heavier so that it wouldn't slide as much, but I don't even know if that's possible in UE4.


t-thanks user

>Ara ara, do you need some pointers on how to please a real woman, user?

The problem is Blender's viewport performance, which is abysmal. Object, sculpting, weight painting, edit, etc. mode work differently.
Edit mode is by far the slowest. I think this has to do with how the model is rendered in it. When in edit mode, Blender doesn't merely render the faces, but also the edges and vertices which means that the amount of shit that has to be rendered gets multiplied.
It's far less efficient that it could be though and that's a known issue. A project to rewrite the entire viewport has been in the works for years and has been slowly integrated piece by piece.
I don't know if the viewport performance in edit mode will get fixed, but 2.8 will have a major viewport update. They are adding a realtime PBR renderer to Blender called Eevee.

So if you're making a game with pixel art how do you ensure everything is the correct size?

Do you know what pixels are

In my opinion SDL is absolutely terrible. I had to drop a project of 2000+ lines of code, because I was so sick of SDL. That even caused me to stop deving for a while - but I ended up coming back, this time using Allegro 5, which is incomparably superior. Choose your tools wisely!

Render to a draw target/buffer/whatever that's the original resolution then scale that up.

Let me guess, Unity ?

Never scale anything that isn't the canvas.

(checked)
Those edges and vertexes are practically rendered as lines and points/cube no matter what shading mode is used, so wouldn't rendering it those stuff still be "cheap"? I mean its not like that those edges/vertex are using some fancy shaders or weird texturing stuff that makes it even more computationally expensive to render it. Though maybe Blender does it a bit differently in this case.

Yeah I could find some few users post on some other websites where they complained about the edit mode performance too and nobody has like a definite answer to it, they only made some observed post about it and compared it a bit to other programs like 3DS max or Maya has better performance for this stuff or so. I have read a bit about their new viewport project but they didn't mentioned much about performance right now.

And they also want to stick to OpenGL 3.3 (core) API too with that version, so maybe some performance improvements can be accomplished that way, right now I don't have much hope for this version but maybe it will change in the future if they are done with those task that they working on right now. Also I forget to mention that somebody told me that Blender uses mainly the CPU when in edit mode and the CPU code somehow doesn't fully support multi-threading/cores which would kinda explain the poor polygon support, which makes me wonder why it is even using the CPU in the first place for this task.

I'm not sure what the proper use of passing arguments by reference is, but is this okay?

UtilityFunctions.pathObstructed (this.transform, ref wallNormal, ref isRunningIntoWall, ref isObstructed); public static void pathObstructed(Transform root, ref Vector3 wallNormalOut, ref bool isRunningIntoWallOut, ref bool isObstructedOut) { RaycastHit hit; if (Physics.Raycast (root.position, root.forward, out hit, 2)) { isRunningIntoWallOut = true; wallNormalOut = hit.normal; isObstructedOut = true; //return hit.normal; } else { isRunningIntoWallOut = false; wallNormalOut = Vector3.zero; isObstructedOut = false; //return Vector3.zero; } }

All of these need to be set at the same time when the raycast finds something in front of the player so I thought it would be a good idea to make it something like this.

Is this good?

yes I know isRunningIntoWallOut is pointless but it's a placeholder for the dot product between my movement vector and the normal of the wall

Looking to maybe pirate Maya because I've been following a dev's modelling videos in Maya and it looks leagues more intuitive and smoother than Blender for modeling. Any thoughts on this, Holla Forums?

Might as well. Also, you should make an account on CGpeers right now, today should be their monthly registration window.

I guess i have to at least modify it, last time i tried to build a CC on my own in bullet i lost all motivation for about two months because it was hell on earth.

Could be just UE, with Unity's pretty smooth

That was just Irrlicht, a pretty barebones game framework in C++, no editor and nothing

For learning you can just use the student version.
On that topic: Does anyone know if there is a chance for Autodesk to find any watermarks in a finished game from ripped assets?

It's forgivable if the rest of your game is written using OpenGL 1. Immediate mode is pretty useful for a map editor (though I'd probably just use OpenGL 3 or 4 the sane way, myself). I thought you were doing something stupid like using a different rendering framework and just using OpenGL's fixed functions for math.

I wasn't saying you should do it with C++. I was saying how I would do it.

That's probably why then, give it a shot you never know

I have to use OpenGL 1.1 anyway because my thinkpads don't support any higher version as well, and i'm using my old thinkpad t41 as an optimization target for the game.

My thinkpad brother. Actually I'm a bit better off, I can at least use 2.1.

When u export a model in obj format it says it's made with maya + version.
Also, if u save a project file in maya w/student version, if u open it in another version, it says "made with student version".
i only use maya for xgen personality, and import back into blender for everything else except sculpting (zbrush).
Also they don't care enough to check, in addition to it being hella bad PR, but it's better to be safe than sorry.

TL;DR
if you erase the maya comments in the exported model file they can't tell, no.

Well, it does make perfect sense. If the underlying data structures don't scale well in terms of parallelization, CPUs are the only way to go because they excel at sequential computing.

Look at the term "pixel perfect".
It covers this, and each engine had their own guide.


I concur

I'd imagine there's a couple factors into it.
Probably one is the huge indie boom we've had over the past several years, though.
Anyone can be a gamedev and anyone can make game. Literal nobodies can become rock stars.

They also hide information in the least significant bits of various numerical values, plus probably other shit we don't know about.
Nah, they threaten you with a lawsuit and usually it's better for you to settle outside of court, which is why you rarely hear about it. This only applies to games that make money, of course.

Try this is you bare using unity.
itch.io/jam/lowrezjam2016/topic/19276/unity-setup-request

I would make it return a bool. More precisely, whether the path is obstructed or not. Then the function's return type would be in line with its name. Then you could also put it into an if-statement directly.

Even though I use Unity, I have to say each has their own issues like this.
Unity has an issue with some SJW employees (beginning to form an echo chamber like Google), and also their recent CEO (he's the former EA CEO); the former started shortly after said CEO started.

Basically, bad angle to take.
For most they're just too invested in one engine to switch to better alternatives (I've got 2k+ hours into my unity framework f.e.), or the alternatives lacks specific features they need.

why would it being settled mean people don't hear about it?

I doubt that. I'll need proof.
As there's numerious problems with doing that.


Terms of a settlement can include stuff like that

Is there a way to rip models from unreal that would preserve these watermarks?
Because if there isnt, shouldnt the only way to get caught be letting anyone gain access to your files or outright idiocy by telling the world you use maya?

There is zero problems with stenography.
Oh sure, I'll just link the handy dandy data sheet that shows all their antipiracy stenography. That stuff surely isn't being kept under wraps or anything.
This is entirely hearsay, but it's hearsay based on relatively simple methods so it's pretty likely that it's true. Or rather, it is so simple that there's no reason for them not to do it. And much like printer dot matrices, it's hard to notice and you can't really prove it is actually anything meaningful without knowing the exact methods by the one hiding information there.


If Unreal can render it, you can get the data out somehow. Epic almost certainly has a tool for it and would provide it or data extracted with it to Autodesk if need be. You are right that they're unlikely to check in the first place if you don't say you use Maya while you don't have a maya account linked to the company you have to set up due to tax shit long before you earn enough money for them to be interested in you.

I meant influx in this particular thread. For some reason there were like 4 different anons in span of few hours asking for tips on how to begin gamedeving.

i dont doubt that they can rip geometry out of the game. but if you want to hide a watermark you would have to hide it somewhere, where it doesnt actually affect anything.
And why would Epic be interested in such a tool existing? If they have it, it can be leaked and ruin the trust of people in the company and not just trust of little guys, but actual businesses, that probably dont want to serve their assets to the internet on a silver platter.

True true.
Better to be safe than sorry, and just import the file to blender; so you have plausible deniability plus potentially removing any hidden messages especially if you change model formats.

For proof I meant more like how it's possible, and examples.
I see how it's possible though, good points.

Godot

...

You wanna keep me in the lobby

Is there room in your love tonight to take me away

You wanna keep me in the foyer

No room for a broken man to pay the way

OK. That can work. The OpenGL 1.x matrix stack is pretty nice.


Nah. Chill out. He's writing a level editor. OpenGL 2 is fine for that. It's true that drivers don't handle the obsoleted stuff well, but it all works in compatibility contexts.

I use GLM, C++, GLFW. It's a pretty nice combo but I think it's actually more complex than the old GL matrix stack. There is just so much more details to take care of. Older GL1 and GL2 "immediate mode" was so nice. I think it's a pretty solid place to start. Everything under the sun is going to run it fine too.

Yeah, in theory you should see a performance improvement when modeling/using Edit Mode with a better GPU.
In practice, OpenGL 2.1's Direct Mode means that all gl commands are executed on the GPU immediately, resulting in a LOT of talking back and forth between the CPU and GPU.
Since in most desktops the GPU is a completely separate brick of hardware, CPU GPU communication is high-bandwidth but high-latency. In Direct Mode, you're constantly stuck waiting on this latency for every gl command.
OpenGL 3.3's more modern API greatly reduces the amount of CPU GPU talking, which should result in very significant performance boosts when using Blender's viewport in Blender 2.8.
However, it's important to note that the process of modeling is something that, by its nature, involves more CPU GPU cross-talk than, say, a video game (where you're more often uploading finalized, static, unchanging data to the GPU for rendering).
Add a new vertex to your model? Well, to render that in the viewport you're going to have to send that new vertex over to the GPU.
Most functions to edit the model are going to be executed on the CPU, too.
So both CPU and GPU performance is important when running a modeling application.
Finally, you didn't specifically mention sculpting, but it's important to recognize that sculpting is an INSANE hit on your computer's resources.
If you don't know how to sculpt in a memory-efficient manner, you'll quickly find Blender and other programs screeching to a halt, running out of GPU memory, because suddenly you're working with more triangles than you can even imagine.
I'm more experienced in poly-by-poly modeling than sculpting, but, for example, I know with sculpting at the very least you have to be making a conscious decision about dynamic retopology vs. just using the Multiresolution modifier depending on your basemesh (or lack of one) and what you're trying to make.
Chances are, if you're not sculpting, you're not going to run into many performance problems with Blender, and if you are sculpting, you are going to run into many performance/resource-hogging problems no matter what program you use. Learn the technicalities of sculpting and it'll save you some trouble.

Personally, with the poly-by-poly approach, I have a character model "scene" (lots of prototypes and test models also sitting in memory) with 12,000 verts/24,000 tris and have no issue using Blender on a GTX 960 and a 7-year old (pre-Sandy Bridge/integrated GPU) i5 CPU.
If you're running into performance problems, you're either doing something far more detailed/advanced than what I can handle, you're sculpting, or otherwise I suspect you may be doing something wrong, especially if you're working on a real-time video game as opposed to a non-realtime animated 3D film or something.

Yes, Cycles is the only renderer that can run on the GPU at the moment in a vanilla Blender install (I'm not familiar with the likes of LuxRay and other external renderers not packaged with Blender).
In Blender 2.8, a real-time PBR renderer called EEVEE is being added, and it will also do rendering on the GPU (and also viewable in the viewport as you're modeling!)

Also: see vid related.
Even the most complicated of Mario's 3D models don't get to be much more than 10,000 tris.
Sonic's models over the years have gotten up closer to 20,000 tris.
Again, I'm having no problem with a scene of 24,000 tris in Blender, and you really don't need more than that for a game's character model, so you should probably be asking yourself some questions if you're up to something like 50,000 or 100,000 tris.

WHATS THE POINT OF SEGMENTING THE LETTERS INTO TRIANGLES IF YOU DONT STRAIGHTEN THE EDGES TO EMULATE FUCKING TRIANGLES

You know the answer.

you both should upgrade your cpus, you are talking as if 24k tris is enough meanwhile i'm here with a way worse gpu (r7 370) and objects that idle in the 120k ~ 500k range with no problem, it doesn't even have to be a god tier cpu either, i have an X4 740K myself and it does the job.
Displaying and editing you final models might not be very taxing but that doesn't mean producing them isn't, for example this is some hair that i'm sculpting, see that grey blob? that is the edit mode with the multires applied (i had to apply it to show you the wire frame, because at that density blender just displays the silhouette)

So considering that Godot is going to have a C# module, I have to ask, is its 3D engine any good?

Currently, not really.

By 3.0, oh yeah baby

You should have banzai buddy narrate that as you walk around for increased a e s t h e t i c

Oh, so that explains why Unity 2017.1 has been so shit. RIP.

Well it looks good. But we'll se when 3.0 comes out.

God, now I don't want to bother with Unity. Guess I should work on models in Blender for the time being.

I have now added the Ctrl-Z "Undo" feature, so you can fix mistakes easily. I'm going to start working on the "Redo" feature shortly.

So are you planning on making Red Sky 2 in this or what?

I wont make "red sky" sequels, but I will make more games in it. I don't know the specifics of the games I will make in the future, just how the engine will be.

I never really programmed before. How easy is it to mod minecraft.

I want to do a few mods before attempting to create a decentralized torrent tracker solution with something like IPFS or zeronet,and I want to see if I could come up with a way to decentralize hosting a private game server.

it drives me up the wall that nobody has really tried such a solution after countless of sites getting shut down.

How easy is it to mod minecraft?*

Pretty easy if you know Java, but you have to work with what the game and Forge give you. If you need something more, then you're getting into bytecode injection.

Well I made the high-poly model for baking only so I thought I didn't need to give much crap about polygon count or so since I won't be using it for anything else.
So it was more of a CPU bottleneck when I tested out a newer GPU?
But that only applies for rendering and not for editing a mesh since I couldn't notice any changes at all when editing a model in Cycles rendering.
Well I very rarely use sculpting myself so I didn't bothered mentioning it, since I am not really in a stage to make a few rock models for a outdoor based map.

Well thanks for the info.

Well my whole model has like 220k triangles and the chassis model has like 97,4k triangles.

Maybe. Hard to say for certain, a simple observation of CPU and GPU % utilization while using Blender could probably give you a good idea of exactly what's going on.
Correct.
Blender Internal is a scanline renderer (much like the old Renderman used in movies like Toy Story was), it was designed for the ancient days when GPU accelerated graphics weren't even a thing, and as such speed was the default (you have to turn on lots of settings in the renderer for more accuracy).
Cycles, on the other hand, is more along the lines of Disney's newer Hyperion renderer, designed with modern GPU hardware in mind and aiming for realism/accuracy over speed. (If you want to get a toon-shading look using Cycles, you have to turn lots of parameters that are on by default, OFF.)
And EEVEE is basically Blender's take on the likes of Unreal Engine 4, a set of rendering techniques Epic Games adapted for real-time use that Disney originally pioneered for use in one of their movies, now known as Physically Based Rendering.
That sounds OK for a modern GPU to handle, but way too many for a single model to be using.
polycount.com/discussion/126662/triangle-counts-for-assets-from-various-videogames
^Take a look at Forza Motorcar 3 on this list (Xbox 360 game, but close enough to a decent modern example.)
172,753 tris…At LOD0.
In other words, at the point where you're using 200k+ tris for a single vehicle model, you're going to need to prepare lower-poly versions of that vehicle model and program a LOD system (or utilize the LOD features in your game engine of choice), or else you're probably going to run into performance issues.

I'm not sure what game engine you're using, but my recommendation would be to just cut the poly count down, WAY down, to, say…30k at least, if not even less than that.

Forgot to mention, Cycles and Hyperion are both raytracing renderers (as opposed to scanline).

are you by the way tank user?

Oh shoot! One very important thing I forgot.
If you are using a version of Blender older than 2.77, make sure you check "VBOs" on under User Preferences -> System.
As of 2.77 this is enabled by default (the OpenGL version requirements are actually bumped up to 3.3 in 2.77 as a result of this change.)
Basically this feature gives you a fair portion of that performance boost that Blender is getting from its transition from OpenGl 2.1's Immediate Mode to OpenGL 3.3 by utilizing 3.3's Vertex Buffer Objects (VBOs).
(A lot of legacy gl calls remain in the code in 2.77/2.78/2.79, but these are being completely removed/re-written in 2.80).

See vid related. I wonder if this isn't your problem.

the gdscript meme is killing me apart, i really miss muh k&r curly braces and ++i

da blyad))) Yes.


I just wanted to use my high polygon model for baking purpose only, I don't want to use this high-polygon model in in-game at all and GZDoom doesn't support any form of LOD for .md3 format, I dunno how exactly it handles the .dmd format which has LOD but it is based on .md2 so it is a bit older stuff. Not to mention that unwrapping a 40/90k polygon model sounds like a nightmare to me. + I cant program for shit thats why I bothered making a mod for Zandronum 3.x in the first place. So my only concern was about Blender mesh editing performance and nothing else, I already know how to make low-polygon models because I am modelling those stuff all the time when I am working in my mod. My models for Zandronum tends to be in about 300-1k polygon range or so and I won't bother increasing it even more. Right now I have a bit more fun working on my Factorio mod since rendered sprites are more forgiving and I like persistent games.

I expressed my concern about this because I wanted to upgrade my GTX 570 for a GTX 970 GPU for a bit better performance when making models or generally when working with Blender, not so much because for games reason, the only demanding games I have right now is the STALKER series. Welp when Blender 2.8 comes out I can test the GTX 970 again from some other guy that has it and see if its now worth it for making models or something like that. Right now I rather prefer upgrading my GPU because that one is easy, its just a swap and thats it. For a newer CPU I would need to buy a newer motherboard for that which is out of my budget range. I have a Jewtel i7-3770K CPU at 3.50 GHz cloak.

I am using Blender v2.78c already so I couldn't find anything VBO related.
Yeah the object mode is still fast so I don't have to worry about it for now.

to the user trying to use blender to circumvent maya's drm
i stumbled through this blender.org/bf/Autodesk_FBX_License.rtf , which means that they are (probably) aware of autodesk's drm and most likely honor it


that processor looks more than capable, actually your setup looks balanced enough for blender, if you can upload your file i could test it on a real toaster as well as my machine

Here you go: mega.nz/#!xdUXxb5a!BR25UspOEtSpTMygsgiw1J2BeHpPgZRywQ0N93VmOQ8

first of all nice model
object mode is fast, it almost made me doubt your claims, on the other hand, object mode is really really slow, like what the fuck is going on, i separated everything by materials and is still slow.

on the other hand separating the cannon and the tires from the body seems to do the trick, now i can use edit mode without wanting to kill myself

Welp so I need to make my models from now on in a more separated ways instead of keeping them together if I don't want any lag while modelling, it sucks that it has to be done that way but if there is no other choices then I guess I am going to do that, thanks for checking it out m8.

You mean the edit mode? Or by moving all the models around in object mode?

Is there any guide to parallax scrolling/backgrounds making?

edit mode, edit mode was really slow, here volafile.org/get/dulTPAKkpmtm/SU-100_HP_export.blend is now separated, i was going to edit it further but i have to sleep, maybe you can decipher some of the changes i made, specially the use of groups and mirrors, by the way, now it works on an 8 years old machine with intel integrated graphics

What do you need a guide for? It's just landscape images that scroll as the camera moves.

on how to make a good one

That's entirely dependent on your artstyle, there are no guides for that.

damn dude, looks like I have to wing this one out.

Thank you user that is some great optimization.

It's mostly just art skills yeah.

There's some basic "rules" though, for example it shouldn't be too distracting or overpowering. Making it desaturated and lowering contrast and lessening the number of different colors are simple ways to do that.

...

i have an artist friend with great potential. he doesn't draw often anymore, i think perhaps he needs some more encouragement.

what do you think of his pixel art? it's not much but it's all i have from him.

My idea for automatic ledge grabbing has failed massively, pic related.

How do I make it so that my character only finds ledges to grab on to, and not making the same happen for walls?
That is without using special colliders?

Or should I be using special colliders? Sounds like it'd be a lot of work and annoyance for no reason

I really like it, you should get him to work with you on a FFT styled game

Nice use of light and shadow, he seems pretty educated. Tell him we think think he's a coolguy.

That's a nice boulder

He has a pretty good grasp on colors and shading.
If he gets discouraged easily, tell him to work on something smaller and more manageable, and to not get too lost in details. A lot of artists are perfectionists me included but it's something you can learn to manage.

I'm a pretty shit programmer so take my advice with caution.
Separate the appropriate sections of movement so that things only happen when they're supposed too. eg. checking for ledges only when the vertical velocity is greater than 0 in a downwards direction or only when you can deduce that you just jumped through player input etc.

So I'm basically casting a ray facing down from the end of the ray I'm casting in front of me

That might actually work, I'll try it out.

very nice visualization btw

Wouldn't it be easier to just add a simple box collider in front of the player while falling and let that collide with ledges? On collision it changes the player to grab state. Seems like that would be less headaches than using raycasts.

So you're saying, spawn an actual box collider in front of the player if the raycast in front of the player isn't obstructed by anything and let physics do their work?

That does sound like it might work, however the player wouldn't be able to snap to the ledge and you would risk having the character hanging on farther away from where the edge is

No, what I mean is have just that collision object. If it collides with whatever you deem grabbable then you snap to the appropriate position and enter the grabbed state. How far away you can grab and leeway for how far below the ledge you can grab are just the depth of the box and how high it is.

It's very similar to that other solution but I think it would save you some headache in needing to determine where you'd need to snap to, since with a ray you'd need to determine your offset from the top face (which may not be the highest part of the model) etc etc.

Just an alternative to think about. It comes down to how you want the ledges to work too. Whether all ledges can be grabbed or only certain parts, etc.

But that would mean also having a bunch of separate colliders on top of the platforms I want to grab. Because there's no way to make it understand if something is a ledge or not, it would just confuse walls for ledges and so on.

Either that or I have no idea what you're trying to tell me

I'm confused user, why would you expect a raycast to automatically know what is a ledge and what's not? With either method you'll need to determine if the object is valid to be grabbed onto.

But I don't, the way the other user sketched it out it's just able to check if something is a ledge without using anything else but the collider of the platform itself.

If I need to determine if the object is grabbable or not I can just assign it a tag and have one of the raycasts check if the tag is right or not.

Like I've said user: whether you use a ray or a box, you still need to do the same things to implement ledge grabbing. I was just suggesting it would be easier with a box because you get to check an area rather than a point would have some convenience in getting the right height to snap to.

My head can't take this going around in circles so I'll leave it at that and try and get some shit done.

I just discovered I can leave Unity running the game, write new code, recompile it and have the changes happen automatically without restarting the game

This is amazing

As of now

wew, although I'm not sure how bad this could be for performance

It's a character controller, you're only going to have one. A few raycasts per frame isn't going to hurt, it's when you start sending raycasts to X amount of objects where X can be a big number in some cases that you would get a performance hit. And even then it'd be better to have that code in your single character than in something which can have a ton of instances itself.

That's good to know

OOGA BOOGA WHERE THEY NEW THRED @ FAM?

we're three pages away, settle in fam, it'll take a while

(checked)
Ad dey page thirdeen mah fello' noble afrika' man.

I have implemented dynamic climbing and used a modified version of the algorithm Valve used for zombies in L4D. (Starts at page 29)
I use UE4 and it works like a charm. I used various optimizations to reduce the amount of capsule traces per frame and it has acceptable performance now. Just don't get any stupid ideas like using box sweeps as "optimization", because checking for overlaps with boxes is significantly more expensive than doing it with capsules. If your player uses a capsule for collision, use capsules for that.

I'm happy to see this is very similar to what I'm going for, means it's probably the right way
Thanks user

I still don't think using capsule casts would make it easier tbh, raycasts seems much less complicated

Alright then

I worked on a simple system for ledge interactions, let me know if you're interested in the mechanics and I'll throw you some autistic sketches

I'd like to see, seems like a good or productive idea to collect all the user's methods for this either way

The problem with raycasts is that you are dealing with volumes and you want to prevent the player from climbing into spaces that he doesn't actually fit in. Using raycasts is a hack. No matter how many casts you do, it will be trivial for myself and other Anons to come up with a piece of level geometry that would produce undesired results.
The problem with casting boxes is that it's much, much slower than casting capsules.

I guess that's also true
fuck

Alright I'll rework it with capsulecasts, can't possibly be that hard right

When you're using either Unity or Unreal (don't know what it looks like with other engines) keep in mind that the capsule cast's start and end point aren't the top and bottom. You take a sphere and move it along a line.

oh, I didn't know, that's probably why I was having problems when checking for ground

Looking back on it the thing that pisses me off the most about this is that the OFFICIAL unity tutorial for this casts a capsule from end to end of the actual collider, which doesn't work

...

cool shit, saved

...

Fixed a few things I forgot

Yo, thanks for these!

I mean, as long as it works, right? R-right?

ass and wip face of another loli I'm making

I personally think there's three stages of quality


If you're not on the first stage then it's good enough for people to see

Looks like an ayyyyy right now.

dat ass tho

Is it possible that this drawray and this raycast have completely different start and direction values?

Because what's being visualised isn't the same as what's actually supposed to happen and I'm very confused

Vector3 topCharacterVector = new Vector3 (root.position.x, p2.y + (charCapsule.radius * root.transform.up.y), root.position.z); if(!Physics.Raycast (topCharacterVector, topCharacterVector + root.transform.forward, out hit, 2) && !isGroundedUtility) { Debug.DrawLine (topCharacterVector, topCharacterVector + (root.transform.forward * 2), Color.red);

nevermind I'm a big fat faggot

...

Literally nobody could possibly gain any information from me explaining that I don't know how to do math

Full-boxy awareness!

It feels nice that I can keep a work routine, but I miss playing.

Set an alarm for work time and play time and actually follow through.

get anything done reward my self with some game time
cant do something decide to take a break with some game time
if i ever get on a roll i end up getting interrupted by stupid shit

Anyone have experience with regular expressions? I currently have this:
(N|\*)(?=O|\*)(O|\*)(?=P|\*)(P|\*)
The goal is to find any series of three sequential characters in NOP, but allowing for asterisks to replace the characters. However, I want it so that it's only a match if there's one or zero asterisks, not more. So acceptable combinations would be NOP, *OP, N*P, or NO*, but not N, *O*, P, or ***. Grouping the whole thing and adding \*? didn't seem to help, but I've never done this before, so I may just be doing it wrong.

Oh, fuck you too, spoilers. That's meant to say "not N``, `O`, ``P, or ```." but with asterisks instead, of course.

That's few enough possibilities to just list off all 4 combinations. I haven't used regex too much, but I'm pretty sure that's the simplest way

Turns out Yo Noid! 2 was made with the Super Character Controller, which I was using before starting to work on my own controller because I couldn't figure out how to set Colliders as Triggers in the SCC.

Huh, maybe it isn't broken afterall?
Weird still

I have to do this with about a hundred of those character chains, so it could get tiring real fast. Plus, there can be characters appended to either side of the chain, such as LNOPM, that I have to figure out how to exclude during detection, too. I was hoping there'd be a way to have the computer do all the heavy lifting, which is why I was looking into regex, but you may be right. I'll look into doing it manually.

I know virtually nothing about regex, but I think this would do what you want?
(?(?=N)((?(?=O)(P|\*)|(?=\*)P))|(?(?=\*)(OP)))
It's using if-else clauses, which might be a bad practice or something, I don't know. Or it might even be gibberish that won't do anything meaningful in the first place since it's written by someone with 10 minutes of regex experience.

...

I mean something like (NOP|\*OP|N\*P|NO\*)

Unless I'm not understanding the question and it's any 3 specific characters, not always NOP, in which case a script would be best. In python, for example:
def find_char_sequence(string_to_search, string_to_find): # generate list of possibilities possible_list = [string_to_find] for c in string_to_find: possible_list.append( string_to_find.replace( c, '*') ) # search through input string for possible in possible_list: if possible in string_to_search: return True # or whatever you want to do
you'd need to change it for more than 1 match in the string, wouldn't be too hard with a while loop, or just running the program again with the input string starting at the point of the first occurrence

awww yiss new icons
also im almost finished adding sounds and music

Can you make a video of this without the character? That's a really good gondola by itself

low effort

you can now bind multiple inputs, you can also bind buttons but they don't have good description names (like axis 1:1 (x axis pushed down or button 2)

new thread