/v/ - Video Games

>5 minutes of progress

>5 hours of unit tests

>>14581856

>>SJWpad++

wut

>>14581838

Try a different font, something more matching the spirit of the game. Otherwise, looks good so far.

>>14581865

If I recall it correctly the author explicitly stated that if one is a Trump voter he shouldn't use his program because of muh xenophobia. Was it a chink that is living in baquette land that wrote it?

>>14581869

well, unity is run by leftist lunatics and i still use it

if it works, there's not much of a point in changing software over views i don't really care about

>>14581869

Who the fuck cares what he thinks?

>Don't use my program because Drumphomph is ebil

God, I want to smack the shit out of anyone who runs their dicksucker about politics like that, where it's not even an issue.

>>14581838

those underlines and transitions are spicy

>>14581867

>Try a different font, something more matching the spirit of the game

surely its too early to make a statement like that when the game currently consists of primitives and mock-up art. Unless keyreal has actually made a definitive decision on the direction hes heading and I just missed the announcement

>>14581922

>>14581867

The 3D models and particles are currently all placeholders, which is why the font looks out of place. I'll begin working on the actual environments soon, and then the UI will look more natural. That's the plan, anyway…

>>14581874

>their dicksucker

That’s a good one

Shilling a new DoD vid by Mike Acton relating to the new unity beta features, pretty quality.

You Unity faggos better eek out the last bit of performance u can from Unity and stop being OOP street shitters

i myself can’t stand the old way, it feels so wrong and unoptimized

>>14582232

but OOP street shitting is my life

i'm still gonna wait for unity to release a tutorial on some empty project for this

>>14581733

Very ugly gun but well modeled.

>>14582303

I don't envy the job they have trying to re-educate all the Pajeets they've attracted.

>>14581733

Look at quixel for texturing, its really easy to use, and has photoshop like layers and tools but in a 3D model context w/all layers of textures being paintable/and or/generatable (albedo, normal, AO, materials, etc).

>>14581928

Agreed

>through dick, unity

>through games, unity

Games supply a context where politics has no place, and is in my view a medium where it’s about fun and nothing else. To be misconstrued as a vehicle for some political agenda is really corrupting what it’s all about; though it does serve as a method of identifying “developers” who don’t really care about games, and are using it as a tool instead of just loving the act of playing games in itself.

>>14582323

>OOP street shitting is my life

Hah, yeah it was my life too, up until a few weeks ago. It’s a hard switch, but once u get the pattern down you’ll have a spine-tingling revelation of just how much better this approach is (I know that sounds pretensious, and I’m a faggot, but it’s true to my experience).

Understandable that u want to wait until it’s released as stable, but now is a good time to make at least a basic prototype to learn it.

I’ve got to rewrite my game in this too… and I see myself totally re-doing a lot of it as I can visualize way better approaches that quickly and efficiently solve the real problems; instead of adding layers of abstraction that are only necessary due to the mental model one builds using a more OOP influence design pattern.

>>14582336

My bet is that the pajeets will stick with the “easy” (I.e. already learned) OOP approach, and not migrate to the new API; creating an easily identifiable division between developers who care about performance/games, and those who are just shitting out games.

>>14581856

Well its open source so he/she/they have no say in what i do with it

>>14581873

Unity has shitty practices backing it

>>14582232

High Performance Python when? :^)

>>14581874

>Who the fuck cares what he thinks?

Well I just wanted to explain the other anon why naming it as SJWpad++ is a suitable option. Since I am a loonix fag I don't really care much about it since I have other text editors/IDE to choose from.

Cookinganon on suicide watch

>>14582502

>STILL adding new features

>>14582467

Wanna bet it will be the same old shitty recipe system every game has, but with autistic detail where it doesn't need?

Like alchemy in kingdom come, a waste of potential making a system with large complexity and shallow depth

>>14581856

>he doesn't separate his data and code

>he doesn't want his data to be easily modifiable

jew

>>14582250

Mike action is terrible to work with, and a massive numale, so he belongs at Unity

>>14582626

Details on him being terrible to work with? Given how shit Unity is, their team probably deserves it.

>>14581856

>not using the source engine

>>14582676

>He expects perfection at the very minimum, while trying to "democratize" game development

After barely putting up with the industry's retarded abuse of OOP for so long, it's understandable he'd have a quick temper. I'm alright with it as long as he pushes good programming practices and people use them (even if it's only to make the man shut up).

>>14582648

Source costs ~$25000-$50000 to use commercially (I forget which figure)

>>14582678

This nigger knows.

>>14582648

Source mapping and lighting is comfy (except for hammer crashing constantly and doing shit like misplacing vertices if you don't have a user-made patch), but I'd imagine sifting through Valve's code would be terrible.

>>14582676

kek, i love the irony

>I know that a lot of the people who work at Unity are even bigger cunts

u sound like you've either: worked within with unity, or, are talking out of your ass; hard to believe as it's "my dad works at nintendo" tier.

Though I've heard similar sentiments posted online from previous unity employees for a select few people; so the part about a few cunts working there isn't a far stretch, and being common in any real world work environment.

>"democratize" game development

you have the quote wrong btw, its:

>democratize data-oriented programming

so, to make data-oriented programming available to everyone, and to thusly acknowledge that unity has shit standards for performance. which i agree with, and anyone with half a brain that's used unity or played a unity game should agree with.

>He expects perfection at the very minimum

>he's actually good at what he does though

good, and that's all that really matters in the end

>>14582771

If it has catgirls and headpatting, >>>/monster/

>>14582678

>Godot-tan is a thing, and of course Rule34 is still in application

Ah, this universe sometimes…

>>14582776

Anyone who has spent any amount of time watching the "meta-game" of Unreal vs Unity knows this. Fucking SJW white-knight types…

>>14582776

https://twitter.com/mike_acton/status/44985126899290112

it's real, wtf man

Any enginedev fags on here might find https://www.gdcvault.com/play/1025331/Inertialization-High-Performance-Animation-Transitions interesting.

>This talk presents a new high-performance approach to animation transitions that the Microsoft team developed for 'Gears of War 4'. The traditional approach to animation transitions involves blending between the old source animation state and the new target animation state; however this can have a significant performance cost, because one must evaluate both the source and target animation states during a transition, effectively doubling the animation evaluation cost. For 'Gears of War 4', Microsoft eliminated these blended transitions altogether and chose to instead handle motion transitions as a post-process with a technique called "inertialization". Eliminating these blended transitions gave them a significant performance boost because they no longer needed to evaluate multiple animation states during the transition. In this talk, David Bollo presents the mathematics behind Microsoft's inertialization technique, for both vector and quaternion values, and show how it can be easily introduced into an existing game engine.

>>14582776

>>14582795

>>14582805

SJW shit aside, I get where he's coming from. The baseline engines and programming techniques newfags reach for are pretty bad and we could save a lot of headaches through making and promoting saner default options.

>>14582813

Are you going to add a catgirl graveyard?

>>14582813

and can you post some of your actual game development instead of promotional art?

>>14582806

That's pretty clever. I wonder how it compares in practice rather than theory.

>>14582771

u nailed the anime style, gj u deserve a heatpat yourself

>>14582822

good points.

thanks for exposing that info man, it's good to know.

>>14582864

>>14582876

The actual panel video is probably paid GDC content. Most of the free videos are either useless social justice stuff or shit like Yoko Taro and a buddy talking for an hour.

>>14582884

He's talking about acting.

>>14582864

>It's used in GoW 4 apparently.

Right, so it's definitely possible to make passable stuff with it, but I'm wondering if it's much different to work with.

Redpill me on Unity

>>14582896

it's worse than unreal, but the license is better

>>14582899

What if I wanted to make a 2d sprite based game?

>>14582771

Will this have haptic feedback support?

>>14582906

you could just use love2d

>>14582906

For a 2D game you might as well use Godot + C++

>>14582906

use Godot

>>14582906

Godot if you want the usual engine IDE bullshit and all its features, LÖVE for something stripped-down but reasonably easy, and Orx for data driven stuff.

>>14582908

Wtf that graphic is worded in a misleading way.

They apparently own 5% of Activision, wich means they are an importatn investor, but they probably dont even get special rights for that.

They hold 40% of Epics stock and it looks like Epics board of directors also has employees on it, so they dont even have 40% of the seats on the board.

You are still giving money to the chinks, but this implies, that they own them outright and they clearly dont.

>>14582890

Is there a division between the two? I'm not retarded enough to actually to go GDC, so I assumed they shove all the actual talks in with the "woman talking about anime and mobile games" talks

I want to make a standalone First Person game like embed related, but don't know where to start with GZDoom.

I found this guide http://www.lofibucket.com/articles/modding_doom.html on making a standalone mod, but I am still confused about what tools and scripting languages (GZDoom supports 3?) I should be using

>>14582976

Yeah, there's a separate page for free and paid content. Free content is mostly presentations stored as .pdf with a couple videos here and there, while paid has more of everything.

>>14582982

Why is there no pirated version?

>>14583042

gee, that doesn't sound very democratized

There used to be a C# repository of the decompiled managed side of unity on github, before (((they))) shoahed it. Can't find it anywhere else. Anyone happen to have saved a copy?

>>14583188

>butt wings

Surprisingly, they quietly released the C# source code some time after the shoah

https://github.com/Unity-Technologies/UnityCsReference

>>14583210

>remember to study anatomy

>links to animeshit

this explains a lot

>>14583217

She is not creepy, she is clumsy and awkward

>>14583335

Anon's life savings won't last forever, we have to get the bully in while we can

>>14583302

>>14583231

She was the only good thing to come out of gen 6

http://www.modernescpp.com/index.php/no-new-new

Are you ready for the next generation of C++? Pointers are now deprecated in C++20

>>14583231

>>14583302

Purple Tomoko is the best.

>>14583438

More reason to use C. Sadly I'm stuck with C++, because I already wrong all the hard parts of my engine with the STL. I will make an effort to get away from C++11 though

http://www.modernescpp.com/index.php/april-fool-s-joke-raw-pointers-removed-from-c

Phew, should've just checked the guy's front page right away.

>>14583618

they joke about this, but people have been lobbying to remove primitive types in Java for decades

>>14583557

Must be C#. If i remember the compiler was optimized for the more common case (single dim array). Multidim arrays should be comparable in speed to jagged ones, but they arent

>>14583522

>>14583515

Explain why one would want to use pointers when you have zero-cost abstractions that provide you with additional safety features (i.e. safer way to iterate over std::array)

(Of course, an internal representation of pointers would have to continue to exist, even if they actually removed pointers.)

>>14582384

>>14582567

>quixel

I can wholeheartedly recommend Substance.

>>14583188

Not bad, but horns > head wings.

Anyone else can't wait for Blender 2.8? I'm not interested in EEVEE as much as I'm interested in multi-object editing. Right now I need to rely on a plugin to do the UV layout for multiple objects onto one texture in a comfortable way.

>>14582881

>Skype

>Photoshop

>Google Chrome

>Steam

>didn't autohide the taskbar

Prepare to be bullied

>>14583640

I enjoy not having to include std bullshit, not breaking compatibility with 90% of all code written previously and available online, and being a grumpy old man who doesn't like those newfangled automadoos and pointers that may or may not have hit the gym.

NESTED PREFABS

>>14583823

>"we are only successful if you're successful"

>we are very, very, very grateful for your continued love and support

>pulls hands to chest

>unity gets huge amounts of money from failed individuals/teams that buy their shit or assets and then fail

>charges for tons of services that cost money regardless of how you do

I've never seen someone who wasn't a Jew lie this blatantly

>>14582776

I both agree and vehemently disagree. Yes, it's very annoying that you have to master programming languages and the various dev tools before you can really make a good game. But at the same time, that need for mastery is what keeps lazy shitstains who just want to make a quick buck making nothing but shitty Unity Personal Edition games using store bought assets.

I'd much rather game dev tools go in the direction where a good dev can one-man-army a good game as opposed to some schmuck making a mediocre game without any prior knowledge.

>>14583682

How does Substance handle art styles that aren't photo-realistic? All their screenshots they use in their promotional materials show photo-realism. How does it handle anime shit?

>>14582771

It's a minor thing, but looking really closely at your model it seems like some of the straps on her outfit should have the outline effect on them but currently don't. In particular I'm looking at the belt/holster on her thigh where the buckle and the part of the belt that sticks out has the outline around it but the part going around doesn't have an outline, so it looks a little odd. Her bandolier probably should have an outline where it crosses her chest as well. I guess the flat colours are the style of the game, but the rifle is kind of a big chunk of brown and maybe a wood grain texture or some extra details on it would help to break it up a bit and make it more readable.

Otherwise she's very cute, and I'm definitely interested in seeing how your game turns out.

>>14582771

Maybe you could give us a few pointers on how to make cute anime shit

>>14583188

>>14583628

Yep, it's C#. I'm trying to learn the internals of it, but there really aren't many resources beyond just looking at the bytecode.

I'm targeting Net 4.6.1 using microsoft's latest C# compiler. The generated IL code for the functions all require the same stack space, but two of these use 8 bytes in instructions and the other two use 11.

        public int OneAdd(int a)
        {
            a = a + 200;
            return a;
        }

        public int TwoAdd(int a)
        {
            a += 200;
            return a;
        }

        public int ThreeAdd(int a)
        {
            int x = (a + 200);
            return x;
        }

        public int FourAdd(int a)
        {
            return (a + 200);
        }

1 and 2 have identical bytecode. 3 and 4 have identical bytecode, which is different from 1 and 2, even though they all effectively do the same thing. That means that ThreeAdd is faster than OneAdd. Fun, right?

>>14584126

That's cel shading (same as >>14582771), although without the outline and with specular highlights. The hair seems to use a different shader though, which is a good thing as the hair is arguably the worst part of the pic from the other Anon. (Hey headpatanon, looking into using a material/shader that posterizes for the hair might be worth looking into. They look and work very similar to cel shading, but don't reduce the color gradients to two colors, but any number you wish. On the screenshot that Anon posted Nintendo used four. To get the specular highlight to look like hair rather than a flat plastic surface, you could probably use a normal map.)

I'm pretty sure I could port one of the million cel shaders available online to Painter. Although I won't make any promises, because I'm busy. I'll look into this some more tomorrow. It's 3 am here.

>almost a quarter of the thread and no one sperging out about racism

>actual developers posting

w-what is this, this hasn't happened in months

>>14584126

Getting good cell shading is largely about getting good normals.

There's a GDC talk that goes into a good amount of detail on how GuiltyGear Xrd pulled off it's art-style. There are some good information in there even if you don't want to make a 1 to 1 copy of it.

>>14584238

Don't jinx it.

>>14584114

>C# compiles to object-oriented assembly

what has god done?

>>14584246

This process is extremely time consuming and unless you can afford a full team I would not recommend it, the results however are excellent and well beyond other games that use cell shading.

>>14584278

The transitions are still soft though. It's also rather intense and I have to say I like the BOTW one better.

>>14584289

That's why I said there was good information even if you don't want to make a 1 to 1 copy. I agree wholeheartedly that It's not viable to do exactly what they did, but some of the stuff they did is viable for a one or two man band. For example the way textures are laid out, using a tint texture for shadows, and maybe even the way they handle internal lines. I don't expect anyone here to control normals to the degree they did, have 600 bone characters, deform the mesh every key-frame, or have separate lighting for each character that changes as the characters animate.

>>14584328

Blend spaces?

Object-Oriented Programming in Assembly Language

By Randall Hyde, March 01, 1990

http://www.drdobbs.com/embedded-systems/object-oriented-programming-in-assembly/184408319

What does an object-oriented assembly-language program look Like? A better question to ask is, "What is the essence of an object-oriented program, and how does one capture it within an assembly language program?" Once you strip away the gloss and notation convenience provided by languages such as C++, you'll find that the two main features of an object-oriented program are polymorphism and inheritance.

me champ

>>14584114

I think it's pretty obvious that compilers do this. Pretty much every compiler is going to have some problems properly optimizing things that seem equivelant due to internal representation. Even C/C++ programmers have to fight their compilers sometimes

Here's another example of similar failure:

I made a StringBuilder wrapper that has a much needed + overload for the Append method (so I can quickly refactor code that uses plain strings to use StringBuilder with little effort.)

The compiler is pretty good about optimizing most away calls and directly calling the Append method of the internal StringBuilder when I use it,

x += y + z + w compiles to a string.Join(…) call and then calls the Append method.

doing x = x + y + z + w correctly compiles to chained Append calls, so I write most of my long string processing that way, and use interpolated strings for other stuff, like printing out messages.

In your case, I bet that an assignment to a parameter is represented differently internally than an assignment to a new variable. The first two reassign a, the other two make a new variable and assign to that. The program tree for #3 probably gets optimized to the same as for #4 due to the very obvious unecessary need for the variable.

>>14583204

I was pretty surprised too, it’ll come in handy for sure

>>14584056

I hope that’s just a winXP theme, as they dropped support for XP.

To get access to the job/ECS functionality you need the “special” beta to get it working (they have 2 versions of 2018.012b, unorganized, I know).

Look at the ECS/Job beta sub-forum on unity’s forum, and go2 the welcome page; u need that beta (2018.012b) linked on that page. Then u need to change the .NET. version to 4.0, reboot the editor, and edit the manifest.json to import the entities and job packages.

>>14583557

Luckily there’s quite a few algorithms for linear-izing multidimensional data.

Look at how compute shader multidimensional threads are linear-ized, and the same algorithm can be applied to data packing.

There’s also Morton codes for cell/neighbor cases.

>>14584568

The main point is the compiler is unpredictable. People routinely and dogmatically insist that the compiler is fantastic at basic optimizations, when it's clear it misses a bunch of basic shit, and fails to optimize increasing variable by 200 and returning it if you DON'T use an extra variable. You then also get to guess on how the language runtime is actually going to implement the opcodes (like a switch statement).

>>14584114

>>14584568

Don't worry about the IL code. No really. Sure its more instructions but unless you're bottlenecked in a performance critical loop, it won't ever make a perceptible difference.

That said, profile it. Also if using VS then make sure the Compiler Optimizations box is ticked under the solution project's properties. They really should look identical.

Also any book by Jon Skeet or Eric Lippert deals with under the hood shit. Jon is a hobby autist and Eric worked on a compiler team.

>>14584664

I'm doing it for fun, I don't expect to write any performance critical code in C#

>>14584679

Implement this for fun, I want to see the IL for this now

>public static void AddFive (ref int value, out ref result)

>{

> result = value + 200;

>}

>>14584728

>>14584728

I can't fucking understand your post. I know it's MC but post relevant snippets somewhere? It should behave like C#, no?

It sounds like whatever calling the code is doing so weirdly

>>14584728

>>14584738

Like why are you making a hard subtype when only the instance data gets changed? Am I missing something or are you just following the pattern it had in place?

>>14584591

Unless you're writing everything in assembly, you don't really know what your program actually is!

Compilers are unfortunately a neccessary evil.

>>14584738

C# and java are similar enough on the surface level that most assume they're the same, and different enough when you deep dive that you realize that they handle very similar things completely differently.

Just a few C# structs are user "java primitives", C# generics are real generics, C# loads entire assemblies instead of individual class files

>>14584694

I think you meant out int result, so thats what I compiled.

not the same guy obviously, but here's the il for it:

https://hastebin.com/ugafaxuhov.go

<it thinks IL is GO

>>14584765

Hey cool, ref/out really does save a few lines of IL for static math functions. Obviously not worth it for general code but it's a nice side optimization for tight loop code to use

Did you compile it with optimizations on? Also what did you use to view IL

>>14584779

I made a new console project in vs and compiled it straight with defaults

and I opened it up with ILSpy version 2.4.0.1963

>>14584765

you nigger, you've compiled that in debug mode, which is pretty useless for comparison purposes

>>14584800

>>14584795

Yeah turn optimizations on, then

>>14584802

>>14584800

whoops, here.

https://hastebin.com/qogupijike.go

>>14584746

No it's a totally new block class and as far as I know it's not a hard subtype. Only the block ID is registered, and based on the metadata it gets a block texture and name but the metadata is never specified anywhere else far as I can figure (metadata/subtype/damage all mean the same thing which further confuses me) but I can make it so that if block:0 is harvested, it drops block:1 and so on and so forth by using damageDropped inside the class

but damageDropped supersedes idDropped for some reason I can't figure

metadata is only ever set by crafting recipes or world building I think, the block classes get that metadata to then do things like change determine damageDropped or texture

	.method public hidebysig static 
		void AddFive (
			int32& 'value',
			[out] int32& result
		) cil managed 
	{
		// Method begins at RVA 0x2342
		// Code size 7 (0x7)
		.maxstack 8

		// result = value + 5;
		IL_0000: ldarg.1
		IL_0001: ldarg.0
		IL_0002: ldind.i4
		IL_0003: ldc.i4.5
		IL_0004: add
		IL_0005: stind.i4
		// (no C# code)
		IL_0006: ret
	} // end of method Sums::AddFive

	.method public hidebysig 
		instance int32 AddFive_Local (
			int32 'value'
		) cil managed 
	{
		// Method begins at RVA 0x234a
		// Code size 4 (0x4)
		.maxstack 8

		// return value + 5;
		IL_0000: ldarg.1
		IL_0001: ldc.i4.5
		IL_0002: add
		// (no C# code)
		IL_0003: ret
	} // end of method Sums::AddFive_Local

So the static add uses 7 bytes of and the local add uses 5. Both have a stack size of 8.

NOTE: All autism here is a work of fiction, any potential speedups are a product of the author's imagination. Any resemblance to actual programming is purely coincidental.

>>14584817

What was the body for AddFive_Local?

>>14584826

Nevermind I'm fucking retarded

>>14584826

I think I screwed that up, they might not be equivalent in function. Why do we have both ref and out?

>>14584832

because that's what the other anon asked for

time for AddSix

>>14584834

>>14584832

ref/out is a pattern that you see when dealing with certain static functions that "do something" like math. For example, XNA has many different addition functions for a Vector3:

>Vector3 v;

>v.Add(other)

>v += other;

>Vector3.Add(v, other)

>Vector3.Add(ref v, ref other, out result)

Anyways, ref/out as pass by reference (rather than value) so it avoids copying the parameter into the method (useful for larger struct types), and just uses it directly. So even if this uses an extra IL instruction, it should be faster. Now I'm curious and want to write my own tests.

>>14584832

>>14584817

Local calls have more call-site overhead.

https://hastebin.com/ozuzobehiw.go

>>14584840

ref by itself is enough in this case (since you can modify the ref you are given)

out is a ref that may not be initialized when the method begins, but must be by the time it returns.

>>14584845

Correct. In this pattern, this usually signals that you're writing to the out value (since it's return value is guaranteed).

Is the anon writing an NES game active? I've been planning one for a few days but I'm having trouble figuring out how to properly organize my code.

Man, one man dev is fucking tiring. I really need to get some buddies with me.

>>14584940

Disregard, I had to hit a refresh button for some reason.

Please be patient, last time I fiddled with this program was in like 2014 to look at Terraria source

>>14584940

ILspy fucks up sometimes, try turning it off and on

>>14584936

Getting buddies is the easy part, the hard part comes after you get drunk and they bite you FUCK YOU FRED

>>14584940

It's not my problem trying to figure out all the new fangled 3D game bullshit. Maybe I should stick to sprite based games.

>>14584805

Here's the custom block class.

public BlockTest(int par1)
	{
		super(par1, 200, Material.glass);
	}
		
	protected int damageDropped(int par1)
	{
		if (par1 == 0) {
			return 1;
		}
		else if (par1 == 1) {
			return 2;
		}
		else if (par1 == 2) {
			return 0;
		}
		else {
			return 0;
		}
	}
	
	protected int idDropped(int par1)
	{
		return Item.coal.shiftedIndex;
	}

So even though I set the item id of what should be dropped on block destruction, I still get it dropping the test block with whatever subtype was specified in damageDropped()

So it's like I can only use one or the other, never both. I even tried setting a variable based on the damageDropped so idDropped could check what subtype was there, but that did fuck all.

Now I looked at BlockRedstone for an idea, and it calls protected ItemStack createStackedBlock() I guess for when silk touch is used, so maybe I can use that with a conditional or something…

What's the least worst scripting language?

>>14584994

scheme

>>14584994

haskell

>>14584994

The one you write yourself.

>>14584994

lua

>>14585032

this guy is trying to trick you into using a non-functional language

>>14585065

You know what a ternary statement is, right? It's basically an embedded if/else statement.

I'm pretty confident that that chunk controls how many item shards are dropped. Eg, redstone or lapis drops 4 fragments. For whatever reason, it still uses lapis as a flag for it? Seriously, try changing it, and I bet you'll get different drops

>>14585065

>notch actually used a ternary operator

holy shit

{return blockID != Block.oreLapis.blockID ? 0 : 4;}
// is just another way of writing
if (blockID!=Block.oreLapis.blockID) {return 0;} else {return 4;}
// I try to keep things as simple as possible when using them, though, so I'd reverse it
return (blockID == Block.oreLapisblockID) ? 4 : 0;

>>14585045

>non-functional language

We aren't talking about C# here.

>>14584845

>>14584940

>>14584946

Okay so I made a self contained program here to test out a few things. I didn't test out the different ways of adding values together (eg += versus a+b), but instead different ways of calling a function:

https://hastebin.com/emagewayap.cs

>>14585110

Back in 2001, a+b actually used to be faster than +=, but they eventually made it so += compiles to a+b. Which was really pathetic, they should have fixed it earlier instead of telling people not to use +=.

>>14585104

I thought you wanted FRIENDS

>>14585120

I heard that FLOPs are faster in C# and managed languages in general, rather than C++, because the former runs in an optimized virtual machine, and in the case of C++, the compiler has literally decades of different IEEE specifications for floating points to account for

>>14585138

I could see some very specifically crafted tests making C# compute faster than C++ (where the CLR/VM compiles it to native code) but I'd be really surprised if C# ever matched it otherwise

>>14585138

No virtual machine can run faster than bare metal. The only way for C# to be better in any sense of the word is for the C++ compiler you using is massively shit and outdated.

>>14585167

what if the virtual machine compiles it to optimized native code?

>>14585171

Then it's not running in a virtual machine, is it?

>>14585171

>virtual macine

>compile

Anon, I don't think you understand what a virtual machine is for.

>>14585181

You write the raw C# code, and then compile it. The compiler changes it into IL code (intermediate language). This is now .NET-compliant code that could have come from C#, VB, whatever.

When you execute the IL, by means of running your dll or exe file, the .NET specfication winds up a virtual machine and runs the code through it, relying on the JITter to hook into machine code on the fly

>>14585181

Sure I do. A virtual machine just runs the instructions of another machine on a host. Basically, its job is just to manage the execution of the bytecode.

In the case of IL (compiled from C#), it knows everything about the bytecode, the functions/variables they represent, etc. Even more, it has access to exactly how they're being used, how long they take, etc. A specific runtime implementation has the right to decide that some section of IL bytecode would be worth compiling to native code, and then it just runs that on subsequent function calls.

This is why the java ==virtual== ==machine== is slow as shit to start, because it's gathering runtime data and compiling/optimizing the java bytecode

>>14585207

u need a new line for red text:

==red== text

red text

>>14585278

Probably a math error on the collision code involving a cross or dot product or normal somewhere

>>14585235

What the fuck just += the literal Vector3.

>>14585278

>>bug appears that causes the player to teleport outside the map and floods the console with a nondescript error message

>I'M FREE!

>>14585292

>anon (1) and done that missed the two dozen posts about tech specs of compiling and benchmarking

>>14585304

Why are you even bothering benchmarking something that gay?

lmao, uppercasing function names

>>14585309

What this guy said. It's not like floats and vectors are important for game development or anything.

>>14585312

I hope that's sarcasm because I'm too tired for this shit

>>14583759

And I enjoy writing fast and safe code.

Have fun with pointer arithmetic, while I enjoy my compile-time static_assert.

>>14585353

I'm not looking forward to C++ becoming any easier, pointers are one of the few things that separate humans who are capable of learning from Pajeets

How would you add into your game a text input box that behaves like the operating system's text input (e.g. has the same shortcuts and copy/cut/paste and everything)?

>>14585438

variable width fonts are NOT your friend

>>14585437

I'm sure you do enjoy those high run-time costs.

Also, if your compiler recompiles everything each time you change something, you're doing something wrong.

The added compile time of compile-time features for a few changed files is negligible.

>>14585446

what

>>14585479

Not using pointers, I mean stuff like checking for some types of errors (compile-time: static_assert), or ensuring you only work with correct units of measurement (like what https://github.com/nholthaus/units is doing).

If you don't use compile-time features you have to do it at run-time.

Using a C array and std::array is the same, performance-wise (zero-cost abstraction).

>>14585453

>compile time bounds checking makes the run time faster

Also, static_assert is literally just reinventing something you can do in the preprocessor:

[/code]

int x = 5;

#if x != 4

#error

#endif

C++ is also missing the bounds checking syntax that C11 has like so:

int foo(int a[static 10]);

[code]

Which causes an error to be raised if the compiler can prove that an improperly sized array was passed into foo on compile time (such as an int g[5]; which is not big enough)

>>14585522

"compile-time bounds checking is faster than run-time bounds checking."

Clear enough now?

Of course C can do SOME things compile-time, too.

int x = 5;
#if x != 4
#error
#endif

a) this looks a lot uglier than

int x = 5;
static_assert(x==4, "x should be 4.")

b) static_assert is not just for bound checking, but for stuff like

static_assert(std::is_unsigned<T>::value, "The type T must be unsigned.");

(or other type_trait stuff)

You can do a lot of template magic with that.

I don't get what you are trying to tell me with your C++ vs C11 comparison.

I am talking about C++11 upwards vs C.

Your example in C++11:

int foo(std::array<int, 3> bar) {
	return 1;
}
int main() {
	return foo({1, 2, 3, 4});
}

This won't compile.

return foo({1, 2, 3});

Will compile.

>>14583438

I don't see why you would want to use the `new` keyword over a unique_ptr/shared_ptr. What's the issue with RAII? Isn't that basically the difference?

>>14585713

RAII is clearly superior.

>>14583640

Faster code, less compile times, less restrictive bullshit, etc.

>>14585604

>"compile-time bounds checking is faster than run-time bounds checking."

I have a very very strong feeling that in 99% of the cases bounds is something you can't even check until runtime. And I'm sure everyone else here does to.

>>14585950

You'd be surprised how far compile time checking gets you when you use it everywhere.

>>14585943

>faster code

What part about "zero-cost abstraction" do you not understand?

>less compile times

Oh, your inherently less-robust code compiles faster. Who cares, with decent design there should be only a few compile units that need recompiling per source code change.

>less restrictive bullshit

As in? What things do you need pointers for, which can't be solved more elegantly and robust at the same speed with modern language features?

>>14585950

Depends on the application, true.

Although I suppose in most cases it's about having a list of things you need to access randomly, right?

Like a list of enemies in your level or something.

In that case, having std::vector and using iterators/for-each-loops is a lot better and more robust than, say, Entity enemyList[MAX_ENEMIES] and then checking if the current enemy is valid or some shit like that.

>>14586010

access randomly/iterate through all

>>14586125

>animu shirt

>doesn't even crop out the original background

This upsets me.

>>14586010

>having std::vector and using iterators/for-each-loops is a lot better

>std::vector

>better

You do realize we're talking about videogames here, right? There's a very good reason multiple actually that EASTL exists, and various alternatives to it in other companies that don't release their version of it.

For reference, one of the optimizations in my todo list is replacing the current "naive" implementation of my SPH simulation using map and vector with one that has specialized containers and search functions.

>>14586156

I am fully aware.

I also think that it is irrelevant for most of the projects here and in many cases just and exercise in premature optimization.

My main point still stands:

I am not arguing for the stdlib implementation specifically, but for the usage of containers like vector, map, array etc. in general as opposed to raw pointers/arrays.

(Of course, if you implement those yourself, you need to use pointers, no way around that)

>>14585604

C11 also has static assert.

>>14586184

You have a point, even I used them for my initial implementation for simplicity's sake and most things aren't as performance-heavy as real time fluid simulation.

I may have a skewed view on these things due to my work with AAA studios, and like I said earlier I'm somewhat like a grumpy old man who doesn't want to learn those newfangled things because they're not how I learned how to do things. Not that I'm actually old, mind you.

On an unrelated note, anyone who parrots that if you'd care about performance, you'd be using assembly instead, should neck themselves

>>14586216

nevermind, apparently you can't serialize objects

>>14586521

Couldn't you make your own simple API with things like CreateBlock, RemoveBlock, SpawnItem, PlaySound, etc, and just slowly, cautiously rewire the native code to hook into it so its not a clusterfuck?

What's the goal of your project again?

>>14586865

oh wait, that only works because there's only 1 block I had that drops the item instead of itself with subtype…

if I have more subtypes and several are supposed to drop an item then returning 0 in damageDropped will do fuck all because then several of them will be dropping the same item

fuck

Does anyone know a reliable workflow to get Real life heigtmap Data into ue4 or at least in heigtmap form?

What im after is picking a specific Area on Google Maps and then getting it into a form that i can use as a heightmap.

Tutorials on Youtube ive found assume the earth is flat and manually measuring distances (or some other mistake on my part) has led to some warping of the proportions.

>>14586958

>muh realism faggots who'll autistically screech about how she'd get Swiss cheesed in a firefight or not be adequately protected from radiation or toxic gasses.

If you get shot with an assault rifle you'd be fucked either way.

>Should I compromise and make outfits that are functionally sound or just say fuck it, go with full waifu materials?

Why not both? Go with Dragon's Dogma style of customization where you can have a slut in bikini armour or play as a heroic knight in believable plate. Have it compact protection a bit but in a way that seems believable.

A girl in a schoolgirl outfit wouldn't be much better protected from gunfire than someone wearing a German paratrooper bonesack.

>>14587006

>compact

Fuck, I meant impact

>>14586958

You already plan to add anime waifus, so going for full muh realism is pointless. I agree with the >>14587006 but I'd add one thing - looks don't have to be realistic but the rules of the world must. For example in cases where there is a radioactive, chemical or fire hazard, just weaning a gas mask on top of bikini armor shouldn't work, it requires too much suspension of disbelief. Character should have to wear a protective suit to survive.

Basically, your job should be to make the clothes realistic enough, but make them look cute, just like in your picture.

>>14587110

I can vouch for this being the correct direction to take.

t. Anon that absolutely does not have a bodysuit fetish.

>>14587095

>>14586958

well depending on the scope of your game, if you have the player constantly reroll/lose equipment. or having it constantly degrade. making the player play with ass-less (damaged) pants or go without. and having proper stalking gear be few and far between.

i think there was a gore system that someone made that used raycastst to delete sections of a mesh via creating an alpha map and modifying the collision so stuff fell out of it. i don't see how half of that cant be made into holy clothing and then adding patch decals on top. when the player "repairs" it.

>>14587192

>I've got nothing but time really to figure it all out.

That dedication is great but those are just outfits/graphics. Where's the GAME

>>14587314

>anon dedicates his life to the cause

>gets hit by a car and dies

>everyone assumes he gave up because it was hard

>>14587451

How much experience do you have with gamedev and compsci shit in general? Ideafaggotry can only get one so far.

>>14585479

You're using a variable timestep for the physics simulation, yes? It's "expected" that it runs at 1/60th of a step every frame, but how do you resolve very large simulations?

Eg, say you had a bunch of input data and wanted to simulate it for a full minute. Obviously you couldn't just do it in one large step, so you'd have to call the update function a number of times, because at each update, a particle might depend on another particle being updated, so how granular do you make the simulation if you get a large time?

>>14586958

Fuck realism, use magic if you need to

Make it fun and enjoyable

>>14587542

That's more experience than most ideafags or I have, which gives you an advantage. Best piece of advice I can give is that making games is fundamentally about designing systems, so approaching your worldbuilding from that angle can help you ascend from glorified ideafag with coding experience to a proper gamedev. This is especially important for a Stalker-inspired game, given how heavily the original relied on its AI autism.

>>14587631

>You're using a variable timestep for the physics simulation, yes?

No. It's currently running as fast as possible (hence the FPS count) while the delta time is set to a static 4 milliseconds.

The simulation is currently not even completely stable, making it good for games comes after that. But even then, having physics run at a fixed timestep is not unusual. I'll have to see how fast I can get it before making decisions like that though. I'm not even sure it'll be feasible at all right now with my own target hardware.

>>14581847

I've been under a rock for a while. Is this a new Unity feature?

>unity's current version of .net doesn't let me suggest inlining

shit

>>14586958

better start a thrwad on >>/agdg/, bro. I wanna see you brainstorm.

>>14588421

>Is there a static variable for ambient light in shaders, by any chance?

there's a global ambient color for the lighting in the whole scene, i change that from gray to pitch black when dealing with daytime systems, not sure about shaders

>>14585740

RAII is worst girl

>testing shit manually sucks, I should write some automated tests

>writing automated tests kind of sucks, I should write a system to make it easier

>that system got a little complicated, better test it

>testing shit manually sucks, I should write some automated tests

>>14587948

Utilize an ECS design pattern for modular systems, and inversion of control (dependency injection) so you can easily switch out systems/components.

For modding the easiest and most effective route is utilization of data driven logic, and harnessing ability to customize entities via switching out and editing component data.

So like kenshi, and fallout/elder scrolls, they use data driven design to allow for easy modding, but to access that functionality each has an external editor.

>>14588481

my bad i meant RAIImu is worst girl

Oi if the anon with knowledge of Ukrainian folklore is still around, I'd love to hear your favourite stories

>>>/agdg/31285

Also other anons, if you want anything included, please chime in

>>14588556

I can imagine you would get that pattern on an RGBA frame buffer by setting every 8th or 16th byte to 255.

>>14588496

SOLID is a pretty good guide on making unit testing easy.

What do you guys do to avoid getting sidetracked by the myriad of distractions out there?

If you listen to music, what music?

>>14588909

in loop

>>14588909

I put my distractions in my game so that my distractions would also be productive.

>>14588889

please make it both

>>14588909

This usually helps me get in the right mood.

Seriously though, I'm working on a dark military themed RPG so listening to the appropriate music can help you mentally. Maybe try finding something related to your project.

>>14588421

>>14588889

Kinda, there's RenderSettings which controls most of what you find in the lighting window, and most of it is piped into shaders. There's an ambient light color defined there.

https://docs.unity3d.com/ScriptReference/RenderSettings.html

If you want your own lighting in your shader (which I'm sure that toon shader already does), you can provide a lighting function for surface shaders, or write your own entire vertex/fragment shader or be a cheating bastard and calculate the same term unity's shaders do for ambient light and then subtract that to undo it, since it's calculated and applied in a part of a surface shader you can't directly control

I don't know how to codeblock but if you inspect your compiled shader you can see how it's calculating environment light. I essentially the same calculation in my vertex shader and subtract it from the output emission channel

It was something like

UNITY_INITIALIZE_OUTPUT(Input, o);

o.wpos = mul(unity_ObjectToWorld, v.vertex);

// The whole point of doing this is to cheat and get access

// To the environment lighting inside the lighting function,

// So it can be nullified during the base pass.

#if defined(UNITY_PASS_FORWARDBASE)

o.envLight = 0.0;

float3 worldNormal = UnityObjectToWorldDir(v.normal);

#ifndef LIGHTMAP_ON

#if UNITY_SHOULD_SAMPLE_SH

float3 sh = ShadeSH9(float4(worldNormal, 1.0));

o.envLight = sh;

#endif

#ifdef VERTEXLIGHT_ON

o.envLight += Shade4PointLights (

unity_4LightPosX0, unity_4LightPosY0, unity_4LightPosZ0,

unity_LightColor[0].rgb, unity_LightColor[1].rgb, unity_LightColor[2].rgb, unity_LightColor[3].rgb,

unity_4LightAtten0, o.wpos, worldNormal );

#endif

#endif

#endif

Probably could be simpler, but I copy/pasted the code from the fully compiled shader. And then at the end in the surface part

#ifdef UNITY_PASS_FORWARDBASE

o.Emission -= IN.envLight * (1.0 - _EnvironmentLight);

#endif

I forget if I tweaked any of this stuff from what unity provided, but I used it so I could have more control over the lighting of individual characters.

>>14589231

not unless you use the same material.

I'd suggest using the same property block across all of the renderers you want to apply the settings to instead of applying individual properties.

I'm determined to murder my cache as much as humanly possible

.Can anyone recommend books/vids for "AI" dev/script. I'm working on a board game. Thanks

>>14589597

depends on how autistic you want to get about the foundation, or if you want to jump straight into it

>>14589597

AI is a diverse topic. Without more details as to what you're trying to achieve the best I can say is to look into Finite State Machines and Behavior Trees.

>>14589659

this guy is your enemy, he's trying to hide all the good information from you

>>14589683

I can't tell if you're joking or if 8chan is fucking up the files, but that's mein kampf.

>>14589671

This is a surprise to me.

>>14588423

I doubt anyone would like a whole thread dedicated to me ideafagging. Plus, I have that natural /agdg/ fear of if I lay out all my plans publicly then someone with more talent (or worse, an actual team) might steal the idea and make the game before I do.

>>14588504

>Thief and Dungeon Siege were designed using an ECS design pattern

Sold, will do it if I can.

>>14583522

>using C++ as a shittier C

Either use C or use modern C++. "C with classes" style programming is honestly one of the most miserable experiences out there. It's not as clean as C and it doesn't provide the same quality of life improvements as modern C++. Note that I don't mean just throw the STL around like mad, but don't be afraid of things that make your life easier with no runtime overhead.

>>14589858

>he doesn't just write in object oriented assembly

It's simultaneously the year 1985 and 2010, get used to it

>>14589858

I avoid modern C++ like the plague because it's bloated as fuck, however I'm stuck with it. I essentially code like it's C but use some of the convenient features like new

So how do you get motivated after a long day of work? Wait until a day off? Get /fit/?

I'm porting my flying game project Hell Divers from UE4 over to Godot, after having put it on hold for awhile. I have a better idea as to how I'm gunna handle the flight mechanics, but I keep over-complicating the physics. Better to start with simple steps and build it up from there, I'm sure. I'm also still fucking around with godot, figuring out how it works. I like it so far, but it's back to square one from the scripting side.

There's so much I want to do, but it all rests on a working flight model, and I have no excuse in not having gotten that outta the way back when I did have plenty of free time.

>>14589883

>modern C++ is bloated

No, C++ is bloated. Modern C++ just makes the bloat worthwhile.

>new

wew

>>14589958

>No, C++ is bloated. Modern C++ just makes the bloat worthwhile.

What does that even mean

>New isn't more convenient than malloc

>>14590028

Looking better, but you still haven't fixed the fucking camera

>>14589989

I mean that it's "bloat" to use C++ when you treat it like C with new instead of malloc. Modern C++ is about low overhead quality of life improvements and compile time features; intentionally avoiding them when you're already using the language is just stubborn.

>>14589946

I have a very hard time staying motivated, my recent boost of motivation actually came from someone telling me that maybe it's time to give up on it, so now I'm compelled to prove them wrong

there's not a whole lot else in this world capable of motivating me like that because there's really no material gain to incentivize me, maybe you'll find a muse or something to incentivize you but it's all down to the individual

getting fit can help your stamina so you have the energy for doing extra curricular things, minimize distractions like social media or playtime, and keep momentum (minimize downtime between dev time, the longer breaks you take between development the harder it will get to build that momentum back up, I find it very difficult to get started again if I've had to take a couple days off from it).

>>14586937

I actually figured out what I had to do for this scenario.

So my revised oreblock class that contains gold/iron/coal/lapis/diamond under one ID needs to return blockID for gold and iron, but coal must return an item ID without a damage value, as does diamond, but lapis returns an item ID with a damage value to say "this is blue dye".

I ran into nothing but trouble with trying to use switch statements for this, maybe I was just confusing myself but I got it to work by having an if statement inside damageDropped() say

if (par1 == 0 || par1 == 1) {return par1;} else if (par1 == 2) {subtype = 1; return 0;} else if (par1 == 3) {subtype = 2; return 4; } else { subtype = 3; return 0; }

so in idDropped I can run a check against this.damageDropped and this.subType, I'm not so sure I like how this is set up though

if (this.damageDropped(par1) == 0 && this.subType == 0 || this.damageDropped == 1) {return blockID; } else { switch (this.subType) {default: return Block.cobblestone.blockID; case 1: return Item.coal.shiftedIndex; case 2: return Item.dyePowder.shiftedIndex; case 3: return Item.diamond.shiftedIndex; } }

I'm kind of annoyed that I have to include that default case, so maybe I'll just substitute it with else if statements

but the important part is that this works, albeit not very straight forward, next is to figure out why quantityDropped isn't working, tried checking against subtype, lapis should be subtype 2 but only drops 1 item, unless I harvest lapis a second time which it then drops a bunch, but if I harvest any other of that ore in the same class after lapis it will also drop a bunch

What do you do when the thing you wanted to make seems impossible on current hardware and you lose all hope?

>>14589177

>Why not just use a normal assertion and turn it off on the release version? Hell I'm pretty sure you could write a macro which does the same thing as static_assert if you want it that badly.

You want it in release, too, though. As said, it costs nothing at runtime.

Also

>using macros

>The STL has tons of un-cache friendly containers such as the maps that group the keys and pairs together for some retarded reason,

You have to know when to use what, in some cases you are better off with a vector or two.

In general though, even when using a std::map the caching won't be your bottleneck unless you are retarded and fill it with actual objects instead of (smart) pointers.

>inheritance which also rapes cache and god knows what else.

Composition over inheritance.

You don't make a custom class for each enemy type, that's shitty design.

>The more I learn about C++ the more I learn about its hidden costs. You'd have to ignore virtually everything else it adds to C.

How about you stop trying to code C with classes?

Treat C++ as a different language, with different paradigms.

C++ does not add to C, C++ is a new language that happens to be able to compile C code.

>1 run-in with boost has taught me never to underestimate how hard compile-time can rape your productivity, and the anon who tried to argue templates don't slow down compile time on here (when it does to a noticeable degree) has taught me to take everything I read disregarding it with a massive grain of salt.

USE PRECOMPILED HEADERS.

Of course everything is slow as fuck if you keep having to recompile the entirety of boost into your compile units.

Yes, templates increase compile time.

But in most cases the biggest source of problems with compile time is retarded design.

Use the Pimpl Idiom, Forward Declarations, Dependency Inversion.

Only include shit you need in the .h, pack the rest into the .cpp

If you have to change one thing and as a result more than one, maybe two files have to recompile, you are doing something wrong.

>Off the top of my head I know void pointers and memcpy are used to put data into an array to be sent as network packets.

Yeah, low level libraries tend to have a C interface, of course you need to use that stuff there.

But if you implement networking, you don't need void pointers or memcpy all the time (except for the very last step, if you only have a C interface to work with).

If I had to do it, I would use a (de)serialization mechanism.

You then can put anything into the network interface, as long as the object knows how to serialize itself into a binary format.

>You can also use void pointers to do generic functions without raping your compile time (like how templates do).

Learn proper software architecture. I'm not even trying to be callous here.

You're filling gunpowder in a pipe and lighting it with a match instead of using a gun, because buying ammo takes too long.

>I am extremely sceptical about the "same speed" part, since GO is the only modern language that wasn't designed for retards in mind. In pretty much every benchmark I've seen C is always the fastest closely followed by GO.

Benchmarks are extremely artificial.

The differences in performance aren't that noticeable in real applications.

Yes, Go scales really well.

Also, about "same speed":

The assembly is literally the same.

If you are interested, here is a talk by Jason Turner, who demonstrates you can code C++17 on a Commodore 64, with almost no overhead.

https://www.youtube.com/watch?v=zBkNBP00wJE

Most modern C++ features are as fast as if you'd do it in assembly.

Modern compilers are MUCH better at optimizing than any of us is.

(Of course, of you do the right high-level descisions. The compiler can't stop you from doing stupid shit in your render loop, for example.)

>>14588909

https://www.youtube.com/watch?v=_W1P7AvV17w

>>14590316

>In general though, even when using a std::map the caching won't be your bottleneck unless you are retarded and fill it with actual objects instead of (smart) pointers.

Even cache friendlier for small sets of objects is storing an array offset as an unsigned char (1 byte) or short (2 bytes), since pointers are generally 4 bytes (32-bit systems) or 8 bytes (64-bit systems). Of course for the char one it's almost useless since 256 items isn't really going to be slow even if every one of those is a cache miss, but at 65536 items it can save a bit of time when needed.

>>14590313

>impossible on current hardware

The hardware is more powerful than you know. Ikusagami, a no-budget PS2 game made by some NEETs in a garage, has a level with 65,535 enemies onscreen simultaneously. The PS3 can do real time raytracing, in software only, using crippled student programming kits. Modern PC hardware shouldn't be a problem.

>>14590806

What I want to do is essentially the equivalent of massive scale (think: 1000*1000+) minecraft block terraforming operations at 30+fps, and I mean more complicated operations than adding/removing items (think: items affecting each other when they move). That in itself isn't a problem, but it becomes complicated since the items need to be split into chunks and each operation may span many chunks at a time, thus each item may need to reference other chunks.

It feels a little more possible after I've given it some more thought, but it still doesn't seem like it'll work out the way I had hoped.

>>14589597 my ip cycled, this is me

>>14589638

>>14589645

>>14589650

Thanks, I'll have a scroll through, I just want to jump into it, learn by doing.

>>14589683

I have that version already

>>14589659

>Finite State Machines and Behavior Trees.

I'll look into behavior trees, I get the idea of state machines and it might be good enough for what I need.

In the game, I have a board, I have 2 players, players drop "pieces" on the board. The pieces come from a dictionary of pieces and player can have 7 pieces in their "hand" before dropping the pieces on the board. I have some old video >>>/agdg/30857 , but will record an up to date version It's not turn based, it's real time. There is logic there to resolve what happens on the board, i.e. resolving when 2 enemy pieces are "fighting", you can play 2 player version of it now, but it's not balanced yet. Some AI would be good to help tweak it and test it.

>>14591060

Sounds like the worst case of optimization autism. It only starts to matter if you're doing it a ridiculous amount of times per frame.

>>14591060

>if statements are heavily related to game dev

ah, so general programming is not really related to if statements..?

>>14591060

Repeating what anons here told me when I asked, unless you have thousands of them, there's little performance impact and using something like switches is moreso a matter of convenience for yourself.

Basically, if you chain ifs in ifs in ifs in ifs many times, that's when it gets retarded. But in your example, there doesn't seem to be too much wrong, unless it ends up being far more complex than you described.

>>14591060

>Is there something wrong with if statements?

Not by themselves. Certain devs get shit for using them in places where a switch would have been more appropriate, and there are places where there's a performance benefit in using the ternary operator instead.

>Say I'm simulating gravity on a player character and I want to render and effect upon landing if the player's falling velocity is greater than some value, is there a better way to do this than with an if statement?

Nah. In general, use if statements when you want different code executed, and consider ternary operators (or similar tricks) when you want slightly different data processed the same way based on some variable.

e.g. if you wanted a different sound effect played based on whether the player is wearing armor, you might be better off writing

playSound(armor ? noiseA : noiseB)

instead of

if(armor)
	playSound(noiseA)
else
	playSound(noiseB)

Unless your compiler is sufficiently clever.

>>14591083

>>14591095

>>14591096

Thanks, anons. This is the impression I had but I was way too much a novice to accept my own opinion.

>>14591113

This is really helpful, thanks a lot. Is MinGW clever?

>>14591060

If statements are fine.

The main problem stems form people coding huge balls of spaghetti code with 100 layers of nesting.

If you have huge if - else if - else if - … branches: use switch statements.

If you have many checks to see if some variables are sane/not null/… consider passing by reference, to avoid some of it, or consider changing your objects so that they can never be in some invalid state.

Refactor tests you have to keep doing into their own functions.

Return early.

>>14591113

>Unless your compiler is sufficiently clever.

It is, in nearly all cases.

>>14591139

MinGW is not a compiler. It is a development environment, so you can use your GNU/Linux compiler (gcc) and other stuff under Windows.

Type

gcc --version

Should be at least 6 or 7.

>>14591139

Oh and btw you can use the Linux Subsystem/Ubuntu under Windows 10 without having to use MinGW.

https://docs.microsoft.com/en-us/windows/wsl/install-win10

So you can develop linux applications under Windows.

CLion supports this, too, I'm unsure about VS, but I think it supports it too in some way.

>>14590376

>using syncthing to sync music from my pc to my phone

>want to save it to my sd card

<yeah no you can't do that since Go doesn't support writing to an sd card and we wrote the android app in Go

>>14591528

Are you sure? I seem to remember that the last dozen threads had plenty of people talking about code. Maybe you just remember visuals more vividly.

>>14591574

he is a "visual learner"

>>14589162

what music would you recommend for a stealth game? where you stalk and grope and kidnap grills

>>14591688

>>14591688

Since people love pushing godot:

Does anyone know a convenient way to get an fbx with a skeleton, materials and animations into godot?

>>14591811

>Na'Tosha

>Assume it's the pink haired abomination

"Hello, World! Well, not exactly. I’ve been around, just quiet. The better part of the last year has been spent on this project: [picture of baby] In early April of this year, Levi and I welcomed Player 3 into our family. Baby Bard is a silly, happy, loud, messy, clever little boy who quickly turned our lives upside down"

>it actually reproduced

>it has a child and still thinks exclusively in funny programming memes xd

>>14591551

>>14591573

These, mainly.

>For the purpose of gamedev, manage to self-teach programming, Photoshop, FL Studio, Audition, Blender, the basics of 3D modeling, 2D and 3D animation, music theory, and writing.

>Find all of them enjoyable to different degrees.

>Need to learn to draw, make textures, and do spritework, but don't get any enjoyment out of them, so every minute spent trying to improve is like pulling teeth.

>Continue to work on game anyway, implement most major mechanics, and write overall plot.

>Game exists, but there's nothing to do in it, because fleshing it out would require being able to create proper assets.

Maybe it'll click one of these days.

>>14591871

Her parents hated her. How many pink haired landwhales can one company possible have?

>I have a tattoo that says, “memento vivere”. It is a reminder of life - a reminder to live. I find myself looking down at my foot and noticing it a bit more these days.