/v/ - Video Games

>>16850041

Here's the proper path for learning programming for gamedev

>learn C++ because you're told it's industry standard

>make a text-based game based on "choices" and hard-coded content

>realize this is fucking awful and start looking for shortcuts

>pick up a really shitty engine and try to make a game that way

>realize you're being confronted with limitations and give up on your project

>instead learn graphical/input libraries for C++ and try to make a simple 2D game from scratch

>wtf am I looking at

>Google "C++ Pong Tutorial"

>manage to make pong in cpp

>try again on the graphical APIs

>make an unplayable mess in C++

>try using an engine

>don't really make good progress

>try making an engine in C++

>no progress

>use the knowledge you gained from making C++ garbage, and then apply that to a standard engine like Unity

>now you can make something half-decent

Oh, and I can't believe I forgot to link this in the OP.

http://gameprogrammingpatterns.com/

>>16850043

Constructive Solid Geometry or mesh boolean as far as i know

>>16850041

>I'm too old it's impossible

Age isn't the problem, you're just retarded. It's ok though, most people here are.

>>16850043

Composite or boolean operations.

>>16850050

Exact same progression I went through, but I went on to GameMaker instead of Unity. Now I just use GameMaker as a library/asset manager and write everything game-logic related from scratch in GML, because the built-in drag-and-drop commands are cancerous. It is a good IDE though.

>>16850043

I believe it's called Composite Solid Boolean Geometry

>>16850053

what are you stuck on?

I don't know much about Godot, really turned off by its scripting language and UI, but I assume it handles things like collision and motion for you?

For bullet hells, you need:

>Object pooling

http://gameprogrammingpatterns.com/object-pool.html

https://github.com/godot-addons/godot-object-pool

Recommended for enemy ships and bullets

>Trigonometry (If you want bullets to shoot out in circular directions from certain points)

https://www.youtube.com/watch?v=jYsG4I2kqvM

Applying the same stuff you learn here should help

>>16850062

He has a point though, learning to code is easier for teens, especially teens already growing up on modern tech.

>>16850062

>Age isn't the problem, you're just retarded.

this. its not that you're old, its just that you don't have the genetic material to do coding.

some people aren't born to be anything higher than being a tradesmen or a retail worker. someone has to do those jobs after all.

>>16850041

I'm fortunate in that I find programming a lot more fun (and a lot less frustrating) than actually playing video games, meaning if I get in the swing of things, it's hard for me to procrastinate. And I can bitch for several hours (or days) but I work on figuring it out, kinda like someone would approach a ridiculous challenge in a game.

That doesn't really mean I'm good at programming, far from it, but I think I'm far better at it than I am at vidya.

>>16850067

Talent is like wearing rollerblades, it just makes you faster, but that doesn't mean you can't just walk if you don't them. Of course there's some who don't have feet at all, and in that case it's probably better to just do something else. Either way the most important ability to have is either motivation or "discipline". All the skill in the world doesn't help you if you spend all your time reading twitter and playing games instead of working on your shit.

>>16850084

It's the effect of living in an echo chamber where everyone pats your ass all day and gives you emotional support and morale boosters. Plus having a religious mindset towards your political alignment, where you feel like you're literally saving the world from evil if you can make other people literally buy your agendas.

>>16850067

The same goes for coders who lack the genetic material for music and art. Thankfully, if you practice under the right conditions your epigenomes for the traits you are lacking will activate and will enable you to do what you couldn't previously.

>>16850084

My sense of pride won't allow me to present garbage, i.e. the result of a nine month coder baby. If they can create something in nine months, imagine what someone who knows what they are doing can do in the same time frame. Chasing that ideal starting as a beginner takes five years.

>>16850085

if you're 25 and you're not already dedicated to coding, you're not gonna get into coding later. its just the way it is. for that matter, at the age of 25 you should already be settled in the field of work that you're going to be dedicated to for the rest of your life. its a pretty good cutoff age for choosing a career.

one of the factors of why is genetic (as in, low intelligence), the rest is maybe its just not what you're really into or simply don't find it appealing even though you have these ideas about a perfect game that you'd want to see.

>>16850093

Looking good, messing with the AI is the best part.

>>16850100

I was an artist with 0 programming experience until my early 20s, and I've never had any problem with programming at all. I had made shitty "games" with multimedia fusion drag and drop shit though.

>what you're really into or simply don't find it appealing

This is the important bit. If you're starting at 25, you should be asking why you never got into it before. Are you sure it's what you actually want to do.

Also wanting to make a game happen is not the same as wanting to actually make it. It's the same as art, admiring great art and wanting to make your own is not the same as wanting to actually draw. The act of creating is not the same as wanting the result, and if you can't enjoy the former then you should seriously reconsider doing it.

>>16850067

No, it could just be old. Try being in your 30s and learning a brand new language unlike the three others you know, or learning a new paradigm like git versus traditional file management.

>>16850489

crop circles when?

>>16850536

When you get our next demo and make them yourself :)

>>16850543

:(

>>16850043

Shouldn't the not sphere be nested under the cube in order to take a bite out of it?

>>16851096

In CSG only the leaves of the tree are primitives. The operations are the hirearchy. The chart posted is bad since it is misleading. The root node is "NOT" , which is above both the sphere and the cube, and then evaluated left-to-right.

>>16850050

Bought K&R C and looked up OpenGL/Win32 documentation after I understood C. Nothing else is needed…. besides math skills.

>>16851300

>Win32 documentation

>he can actually create a windows window in c without copy pasting the 200 lines of code

You have my respect bro.

>>16850050

Most gamedev happens with scripting languages inside an engine that handles all the complicated stuff.

You could use Godot or Love2d

>>16851300

>>16851315

Win32 documentation pisses me off sometimes. It's often times easy to find a variable or function for something but not an example of how to use it. For example, just getting values relating to the primary monitor is easy and self-explanatory. But enumerating and getting info for several monitors is a convoluted several step process. It also doesn't help that their default "lowest supported Windows version" just says Windows 2000, regardless of whether that's true or not. For example, multi-monitor support was added in Windows 98 and most of the multi-monitor functionality still used today works on it just fine.

The dumbest thing I've seen is that Microsoft has deprecated an old function used to get the version of Windows, making it so that you can't recognize whether a computer is running 8.1 or 10 unless you build your application on 8.1 or 10. I saw that 8.1 has a workaround but 10 obfuscates it in several ways.

Anyone know how to mask a sprite3d in godot? Or just to take a piece off of a sprite or something like that?

>>16851388

I've heard that learning to draw (that draw a box website) and studying medical anatomy are good ways to improve 3d modelling by improving spacial and observational awareness.

>be fed up with shitty gameplay in eroge

>decide to make my own

>the more work goes into design and code the more it feels like I'm wasting effort and should just make it a normal game instead

God fucking dammit.

>>16851456

Honestly for and indie dev theres probably more money to be made in making eroge games than regular games.

>>16851459

Set up a patreon or something and you're likely good to go.

>>16850041

at 27 you still have plenty of "neuro-plasticity", most people stop learning new things at 30 only because they don't have the time.

start learning a smaller more interactive language alongside C++. (literally any other language, eg. lua, scheme, smalltalk, python, C).

It is honestly quicker to learn C++ by learning core programming concepts through more educational oriented languages, then going back to C++ when you have some deeper understanding of what it provides you with.

The most important thing to learn as a beginner is abstract data structures.

Learning about lists, trees, stacks, queues, and graphs will turn you from a clueless monkey into someone with intuition and an understanding of abstraction.

to create a linked list in C++, not only do you need to understand the OOP model, you need to understand static typing, template programming and memory management, not to mention all the syntactic sugar C++ has to confuse beginners.

Learning programming through C++ is like trying to study philosophy by first learning Latin or Greek.

The C++ community generally misleads newcomers into cargo-cult mentality.

C++ makes any compromise possible in the favour of "production" rather than the user, which is a significant handicap in learning.

>>16851517

and I should warn you, spending enough time on a computer to master C++ might turn you into our friend "April" in the vid I posted.

>>16851517

>more interactive language than C++ […] e.g. C

Nigger what. Sound advice aside from that.

>>16851517

legit advise

since actual universities that move 20 years behind technology started teaching C++ in 2nd year of CS instead of first, you can make your own conclusions..

>>16851315

You're only writing 200 lines of code because you're following a tutorial. I can do it in 100 lines of code to get OpenGL up, less for software rendering. But still I will copy paste out of my old projects to save time, even though I can do it from memory.

>>16851517

C++ is generally unnecessary to learn and rots the brain… too complicated of a language to fully understand, and too many libraries lets people get away with having gaps in their understanding. Generally locks people into the mindset C++ creates- preventing them from going further. It's like a less restrictive Java or C#, but still a language designed to manage you, if it wasn't too weighed down by its own complexity to enforce this as well as it could.

Depending on the level of patience languages like lua, C, assembly should be used instead.

>>16851902

What do you do to connect them to each other?

>>16851913

A recursive method that "tile by tile" attempts to get closer to its endpoint by checking the angle between its current spot and the end point and picking the best direction. If that direction overlaps with a room or another hallway or goes out of bounds it then goes back and picks another direction and so on.

What you see in those pictures is the same method but the float that it uses to decide its direction is cast to an int, thus the L shape of the hallways. The L shaped hallways all connect fine.

>>16851940

Have you ever thought about something more expeditive like you take a point in each room and just connect it to the closest one by casting a ray and using DDA? Or maybe this:

https://www.reddit.com/r/roguelikedev/comments/1dopkl/my_procedural_dungeon_generation_algorithm/

https://en.wikipedia.org/wiki/Delaunay_triangulation

>>16851940

>I just want something that results in a Z shape or something.

Then, why not space your rooms a little more and connect them with a random walk or maybe combine your rooms with cellular automata?

>>16851988

Tell me more about this random walk you speak of.

>>16852012

Remove the walls for the hallway then, and also, you could try to make your rooms not so angular. Dungeons often have circular rooms and rooms with "diagonal" angles rather than 90° angles.

>>16852016

The problem then is programmatically figuring out what the outer extents of the hallway blob is and only putting walls on that since I don't want players walking off and falling into the abyss. Also having all of those individual sprites and collision shapes for each tile can actually manage to knock down the fps a bit.

Circular rooms and so on will come later. Eventually I want to have very interesting dungeon generation with rooms and caves of all shapes and overhead hallways that go through rooms and so on, I mean thats half of the reason why I chose to make a first person roguelike in the first place.

>>16852031

>The problem then is programmatically figuring out what the outer extents of the hallway blob is and only putting walls on that

I'm not really familiar with 3D roguelikes but if I was in that situation I'd put the floors first then raise the sides who are adjacent to the void, that might do it.

>>16852031

What if you picked points at random on the room walls that face the other room in the pair, extend hallways cell by cell cardinally toward the other room from both of the rooms, then when they are aligned start extending them cell by cell towards one another? You ought to be able to guarantee Z-shaped halls that way, and the wall problem is simplified because you just don't place walls in the cardinal direction you're going except one time for each hallway when it bends. You could even introduce some noise by skipping a step with one or the other hallway with low probability. Also, for your room blob wall removal problem, if you could detect when two hallway cells' walls would overlap, those are the only walls you'd need to remove. In fact, if you can, you could change the wall drawing function to toggle whether a given cell needs a given wall or not. I have no idea if that would work in your implementation though–ideas are easy to spout off on paper like this when, like me, a person knows nothing about the actual implementation.

>>16852034

>I'm not really familiar with 3D roguelikes

Lol thats fine, there basically isn't any.

>put the floors first then raise the sides who are adjacent to the void

Off the top of my head I would have to have every hallway tile raycast out in each direction to know which sides don't have another hallway tile or have a representation of the hallway in a 2d array and do some checks in there. I wouldn't mind doing something like that but like I said Godot starts to chug when you have every tile be its own individual thing with its own collision shapes and such. For the sake of efficiency whenever there are hallway straightaways I tile the sprite and scale the collision shapes to match the length of it. I do something similar for room walls that have gaps in them by breaking the wall down into smaller rectangles and then assembling them around the gaps in the wall.

>>16852046

Yeah that could work actually. Like you said it would only work for rooms that face each other and don't have anything in between them but I think I'm going to try it anyway.

>>16852048

There's a couple actually, Delver and Eldritch. The first one is cartoony and was inspired by Prelude Of The Chambered by Notch and the second is centered around the Lovecraft stuff.

>I would have to have every hallway tile raycast out in each direction to know which sides don't have another hallway tile or have a representation of the hallway in a 2d array and do some checks in there

Why would you do all of that raycast maths when you can just check in a 3 by 3 area which tile is next to the void and which isn't?

>>16852061

Yeah but they're all action roguelikes or roguelites. I'm going for a true roguelike as much as I can.

> just check in a 3 by 3 area which tile is next to the void and which isn't?

How would I do that? What would I check for exactly?

AmDev would be a great boon to gaming if it weren't for one thing…there isn't any money in it.

If there is one thing that the mainstream devs are demonstrating it is that too few anons are willing to pay too few dollars to make developing for console or PC worthwhile. Apps are where the money is, and that means small, short, "pay-per-chapter" design. Meaning they are already heading toward amateur developing, in an effort to corner the market by paying individuals to dev games as 1099 independent contractors rather than as actual employees. You pay for your own insurance, get no benefits, and by the time the tax man is through with you, you need 3 of those jobs to make any money…meaning you are working 60 hours a week. But there are enough anons willing to do it thinking that they will make a name for themselves, or just being "hopeful" that it will work out. But they are being played just like everyone else.

I still want to make a Super Metroid room generator. My problem is that I can easily make a bunch of connected nodes that say "Room B connects to ACDE, Room A connects to BC" etc but I'm not sure how I planarize the node positions onto a 2D plane and come up with realistic room sizes.

This is a general problem, not specific to SM. For example, a room with 3 exits could be one tile in size, or it could be a 6x2 room with a bunch of features and I'm not sure how to design that into a generator. Random ranges and drunk walk dont feel appropriate.

Also the placement of the room in the area itself

>>16852099

Could you show some video example?

>>16852104

So your problem is fixed?

>>16852104

The only thing that gives me any real joy any more is seeing people with problems get solutions to them. I'm happy.

>>16852119

Good job, what are you going to do next?

>>16851388

Consider using software that lets you design characters for use in 3d, you're miles away from having passable sculpting skills, and sculpting skills are a start, you need to retopologize, uv unwrap, bake/texture, rig and animate characters and each of these are their own unique discipline.

>>16852121

I still have a bit of tinkering left to do with old school style dungeons. I need hallways that intersect during generation create proper intersection points. I already have an Idea for how I'm going to do that so it shouldn't take to long. Then adjust my algorithms to deal with wider or taller hallways. Then I go vertical with ramps, multi floor rooms, and obviously multiple floors.

After that comes inventory, combat, spells, monsters, dungeoneering mechanics, and so on. I've been building up ideas for that stuff so I should be ready to go when I get there.

After that probably alternate dungeon generation methods. I want to take the random dungeon generation method out of the AD&D dungeon masters guide and give that a shot. That might be difficult though considering it has diagonal hallways and some nonrectangular rooms and stuff but I have a few ideas on how do deal with that.

>>16852129

Really looks the part.

>>16852132

Oh actually before the alternate dungeon generation methods I was thinking of having a small town outside of the dungeon with stores and stuff like in Moria or Angband. Ive heard that in Shiren the Wanderer the more you played and died in the dungeon small stories would happen in town and stuff and I kind of like that idea so I might take it.

>>16852129

It looks really good, what kind of game are you making?

>>16852132

>>16852138

Good luck with your game, I'm sure you'll make it if you keep going hard at it. Start hanging out in some gamedev forums to help you stick to it and maybe get a fanbase.

Just in case you didn't know this existed:

http://www.roguebasin.com/index.php?title=Main_Page

>>16852142

Thanks man and yeah I've definitely been going to roguebasin. Thats why I've already planned out how I'm going to deal with a polymorph self spell and polymorph traps.

>>16852129

Man, that vibe I'm getting makes me want to go play Brigador again.

>>16852308

its breddy gud

>>16852308

It looks really cool. Only problem is the targetting. It's so small and minuscule you can't tell what you're going to hit and it also makes the whole thing look weak.

>>16852969

Are the rooms going to be shaped randomly?

>>16853066

Yes, that's what I'm trying to figure out. I don't want to just say it's rand(xy) size, nor do a smallest path thing to join them together, because vanilla SM has so oblong rooms once in a while

>>16853075

How important is it that you start with a list and build a level from that as opposed to building a random level and making a list out of it? Because honestly doing it the latter way might be quite a bit easier.

>>16853078

I'd say it's pretty important to generate a room list first. When I have a forest of nodes in the first part, I can attach path-level features to each node such as "AB requires Speed Booster", and then when I actually transform the rooms into tiles, I can take those obstacles into account, plus it helps ensure that the run and items found is valid and makes sense.

But even with a random level approach, I'm not sure of the best way to shape the rooms.

>>16853090

Metroid has a pretty limited number of room-shape types, you can probably just define those in lists and pick a random one based on number of connections and size.

>>16853090

Well for building random rooms you could just over lap rectangles of different sizes and and use the combined shape or start with a room that is one screen and then add a screen onto a random side of that room, going until you have hit a maximum room size or whatever.

As for random levels having power up barriers I don't think that would a huge issue. All you need to know is the order that a player should collect power ups. After generating a level, pick a big room to be the start and then grab a handful of the closest rooms to that one. Put some anti-grav boots in one of those rooms and make anti-grav boots necessary to traverse rooms outside of those original few. Then radiate out from that original bunch of rooms to a new set and put a new power up within those ones, mark that the the rooms still left over require that power up and so on. Something like that might work.

>>16852610

That's just because the targeting is the default ORA airstrike UI graphics, which are at a much, much lower resolution, so they look super small. For the final version I'll just have the notifications be bigger in general. Also more recent versions of ORA allow you to select the direction of the airstrike, so I'll include that.

>>16852093

Not wrong.

The other factor that you didn't mention is that as the major companies downsize, the market is going to be full of actual professional developers becoming independent, and that will give amateur developers a shitload of incredibly stiff competition for those already scarce dollars.

Hey everyone.

Hoping to be a yesdev before the end of the year.

Thinking ahead to the upcoming release, considering some different possibilities, let me know what you think.

1. Post your progress here on 8chan because you have an undying love for anon (only to maybe later get driven away as a yesdev because now you're well-known and are easily identified by your game), use itch.io, don't use Paypal to cash out on itch.io since they'll probably cancel your account sooner or later, maybe use riot.im for private communications, and…I don't know, maybe shill on Reddit/Twitter/Youtube/etc. with mirrors for videos on Bitchute (or as a torrent, or something) for when your videos get labeled as "for kids" and/or get banned?

2. Don't post your game here because it's not worth the risk.

>>16853284

There were successful games made here and their devs surely enough didn't get shot.

>>16853284

Just skip this useless "he he I love imagebeards" phase and go straight to twitter and various google platforms. Imageboards aren't useful for anything productive, and you'll end up moving to mainstream social media anyway if you release a game. You also don't have to shit the thread up with this blackpill cancer about how touching an imageboard will ruin your entire career for life.

>>16853484

I'm on year 7 or 8 with no playable demo.

>>16853504

>>16853500

Oh, so that's the struggle? I see…

>>16853284

see

>>16853304

You can post here and on twatter/faceberg/steam/whaterver. Just don't talk about 8ch outside 8ch and nobody will care.

>>16853871

If I remember correctly, transform.translate is beneficial since you can do localspace translations, as in telling things to go forward and they'll go in the direction they face.

It's literally a function that checks what kind of translation you're trying to do, then either does exactly what you're doing or does the localspace equivalent. Technically doing it directly should b a tiny bit faster due to overhead, though I'm not sure that's true in practice due to automatic optimisations.

>>16853902

If has an overload that lets you set what space it's transforming in. Depending on how well the relativity setting works I may use it more due to some ideas I had for ways enemies could move.

>>16853871

I know I have to expect what I can do, but that's just not nearly enough.

https://itch.io/jam/7drl-challenge-2020

If I succeed in making a game in 7 days does that technically make me a yesdev?

>>16854017

If it's a "full" game that's publicly available, it's yesdeving.

>>16854233

>first person 3d roguelike

Interesting, I considered 3D (on the gameplay level) but decided against it, so I definitely want to see where this is going. Don't you nodev out on me.

>>16850041

25yo here. if you're trying to learn C++ i recomend learncpp.com . i've been using it and learnt a lot, though it might be a bit overkill for making vidya, at least if you want to work with someone elses engine.

question: does anyone here work with Unreal Engine 4? i've got some noob questions but don't want to look stupid in the forums

>>16855126

I've made a few small demo type things. Nothing major. So maybe I can help, maybe not. Lay it on me.

I'll ask a retarded question because finding good resources on this is a pain in the ass.

In C++/Winapi, you can't draw to the window at the same time as the window is receiving messages (for example, being moved or resized). As far as I've read, the solution to this is multithreading, putting the window messages in one thread and the drawing in another. Assuming this is right (if there are other ways to do it, I'd love to know), what is a good way to multithread in the main game loop?

As far as I understand, the function specified to the thread only runs upon thread creation but creating the thread repeatedly should be very wasteful on performance, especially with how often I need this function to run. What's a good way to approach this?

I can actually fix my particular issue by putting the draw code in the callbacks for window movement and window resizing but that seems like the stupidest fucking fix I can think of and I'd rather do it "properly" if at all possible.

>>16855510

You can just have the other thread on standby. Create an event object with CreateEventA on your thread and then signal it using a SetEvent call on WM_ENTERSIZEMOVE. Meanwhile hold the other thread in a blocking WaitForSingleObjectEx. Then execute your game loop in the other thread, checking a nonblocking call to WaitForSingleObjectEx, and call SetEvent again once you get the WM_EXITSIZEMOVE. I haven't done it this way myself but it should work.

Alternatively you can just use SetTimer on WM_ENTERSIZEMOVE to make windows spam you with WM_TIMER messages, which can then be used to execute your main game loop. Just store a pointer to all of the state needed with SetWindowLongPtr/GetWindowLongPtr and call your game loop as a function. This only requires one thread to do.

>>16853099

>>16853104

Having thought about it a bit more, I'm leaning towards a biased random walk. After N steps, the walk ends, and then I can scale by like, 4 or something, and treat each visited cell as a tile to use.

One advantage of doing this is that if I also create non-uniform boundaries (for example, a room HAS to fit within a certain region) I can blacklist certain tiles from the walk

>>16854583

Thanks and how come you didn't go with 3d?

>>16853484

>It doesn't look good enough, no matter how much time I take making it.

graphics or the game in general? unless you're also a good artists it will probably look like crap.

if it's the former figure out what you actually wanna do, worst case treat it as a learning project and do something else in the meantime and come back to it later (or not).

>>16853529

plus some random unknown indie dev isn't worth to sick the twitter mob on you. it's really only something to worry about if your game gets so big on it's own that they feel the need to cancel you if you're not an ally.

>>16855533

That second solution is mostly what I had in mind myself.

Now what about getting other window messages while resizing/dragging? Seems like it should be added in a queue but inputs get dropped if they happen while a window gets dragged. I have callback events for keys getting pressed and released and they get dropped, meaning a key that got pressed normally, then released when the window was being dragged would still consider itself pressed, and rechecking the keys after window movement, while possible, is again inefficient.

>>16855870

Considering I already have a window movement function set, I might just throw GetKeyState in there to rescan the mapped keys after window movement. It's not ideal but better than running GetKeyState constantly. The only time I really need to keep track of ALL keys is during key rebinding which is an obscure enough case and I only need the key being pressed once, not anything more out of that.

Just have to also wipe inputs for any case of the window losing focus too. It always reminds me of playing vidya a decade ago and actually figuring out that you can keep the inputs locked by pressing alt while the key is held. I don't remember what the game even was but I'm pretty sure it was a AAA release. Always found that to be a fascinating oversight.

Last question, super quick one.

I put simple WM_KEYDOWN and WM_KEYUP cases in my windowproc as a test, the only thing they do is cout the key being pressed/released. WM_KEYDOWN starts repeating the inputs after about a second of being held, but if I try to hold two keys at once, it only spits out the repeat of one of the two. I assume this is just a limitation of Windows itself? It does the same thing in any text processor I try as well.

>>16855884

Yeah this is simply windows's default behavior. I've observed it as well. I would recommend using the raw input API if possible, since it's supposed to take any windows quirks out of the equation, alternatively I suppose it's easy enough to use some kind of global boolean to enforce a rule that a WM_KEYDOWN requires a corresponding WM_KEYUP, otherwise you can discard it.

The alt key locks my window and I never looked into it that deeply. Kind of annoying and I end up forgetting about this constantly.

>>16855510

>I'll ask a retarded question because finding good resources on this is a pain in the ass.

It sure as hell is. You shouldn't expect proper answers here; seek out documentation to back stuff up. Every forum post I've found is how the user personally perceives it; but with undefined behavior it can work one day and break the other.

As far as I remember, a window (hwnd) and its associated device context (hdc) are considered the same objects. The OpenGL context (hglrc) can draw onto the window associated with the device context it is created with. These are all written for single-threading, so writing to the window with OpenGL whilst querying messages is technically not allowed.

That said, drawing to a frame-buffer (as you might already do as part of deferred rendering) is different. You don't draw to the window, so it's okay. This distinction also makes sense since you can share a single OpenGL context between multiple windows. It means you can render to a frame-buffer, and when it's time to blit to the window, send a message to the main thread to briefly take over the blitting work. AFAIK, this is the way you can multi-thread it without stepping on undefined behavior.

If you look around for examples though; you'll find something different. A lot of projects out there actually query the window whilst drawing to it, but pause the rendering thread when something like a window move or resize occurs. I believe this is technically not allowed, but "just werks" because internally in Windows and graphics drivers you're (probably) not causing any race conditions. I wouldn't be surprised if Windows developers and graphics driver developers make sure this continues to work in the future for the sake of backwards compatibility, but refuse to document it because that's not what you're supposed to do.

Realistically though, the easiest solution is to just single-thread all of your OS stuff so you don't have to deal with this headache. This might seem inefficient, but there is a fair bit of overhead involved in having threads communicate with each-other, blocking and having to wake up. The main reason you might want to move drawing to a different thread is because issuing drawing commands is prohibitively expensive. This is only the case if you draw an absolutely massive amount of objects like modern AAA games do. At that point; you're probably best off using Vulkan; dispatching helper threads to assist in generating draw calls, directed by the main thread.

I myself have decided to not deal with any of this and just use SDL, keeping things simple. I have spent enough time tinkering a broken ass engines and diving into this poorly documented windows API hell.

>>16855919

Sorry but you're misunderstanding the nature of this question. What you're observing here:

>A lot of projects out there actually query the window whilst drawing to it, but pause the rendering thread when something like a window move or resize occurs…

Is the behavior that >>16855919 is trying to avoid. This isn't a conscious decision but rather a problem in the design of windows- it treats window dragging like a special case and pauses the application thread until dragging is complete. This isn't a race condition of any kind. OpenGL and SwapBuffers work in the callback so it's possible to solve it in a single thread. It becomes more complicated with a multithreaded setup as OpenGL commands can only be submitted on one thread.

>I myself have decided to not deal with any of this and just use SDL, keeping things simple.

This doesn't solve the problem, though. SDL exhibits the same freezing behavior that he's trying to solve. You can find similar issues with every other abstraction layer above winapi.

https://bugzilla.libsdl.org/show_bug.cgi?id=2077

>>16852129

Has anybody ever told you how much of a Fallout/Underrail vibe these models have?

>>16855976

Game ruined

>>16855633

I generally expected it to be a pain in the ass to work with, but especially dungeon generation worried me because I want to have different dungeon "archetypes". For instance, generating a cave is commonly done with random walks, but 3D random walks have very different behavior compared to 2D walks.

I found out that if you make a demo and release your game on steam with it, it gives you almost 100k views since there are autists who download and play every demo

>>16856110

Well yeah. Demos help sales, that's common industry knowledge. Main reason you don't see lots of AAA demos is because

>They already have plenty of marketing and reputation, so people will buy it anyway

>Lots of AAA games play the same these days, so you're not really helping anything by letting people see how the game plays.

If you're just starting out then by all means include a 30 through 1 hr demo of your game.

>>16856092

The function definition inside the class also needs to be conditional to the preprocessor ifdef

>>16856092

Let me say it again: you probably need to conditionally DECLARE the function inside the class. You do this with a preprocessor directive like you did with that function: ifdef.

I think you don't know how the preprocessor works. It's really simple, look it up

>>16856110

A good demo can make your game successful almost singlehandedly, even with no marketing or "influencer" shilling.

>>16856307

>game

No, that's an art gallery.

>>16856314

pics are broken, the code is:

#ifdef WITH_EDITOR

void PostEditChangeProperty();

#endif

i get the error: "Function definition for 'PostEditChangeProperty' not found

>>16856338

You're trying to use PostEditChangeProperty(FpropertyChangedEvent&), why would you declase PostEditChangeProperty() then?

>>16856341

probably because i'm retarded. my logic is that PostEditChangeProperty is what compiler-kun is yelling at me about so that's what i need to declare for it to compile

>>16856355

>>16856341

i finally figured it out. i needed to declare all of PostEditChangeProperty(FPropertyChangedEvent& PropertyChangedEvent) rather than just PostEditChangeProperty

>>16855533

Trying out the second method worked quite well with one small problem - the window freezes for about a second when grabbed and not moved. If it's moved, it immediately starts animating but if I just hold it without moving, it takes a second. I assume the window just takes a second to realize what's going on and to send the WM_ENTERSIZEMOVE message. Wonder if there's some way to fix that, probably using WM_NCLBUTTONDOWN or something though a quick attempt with that didn't pan out.

>>16856045

>a pain in the ass to work with, but especially dungeon generation

You were right lol. I do plan on caverns and things implemented eventually though.

>>16856484

I haven't solved this problem either. If you do figure out a way around this sometime, I'd like to know.

Here's a strange issue I discovered, more a hardware issue on my part than a code issue most likely.

My motherboard supports using onboard graphics alongside a graphics card, meaning I can plug monitors into both. For cable convenience, that's what I have in my setup (had to plug my second monitor into onboard to be able to hook up consoles into my main monitor's second port).

This presents a strange issue where my program, now with disabled Vsync, still Vsyncs itself when it's on the onboard monitor. Other two monitors are fine, and hooking up all three monitors to the graphics card makes them all behave normally. It's either a motherboard issue or an issue with shitty Intel drivers, most likely the latter. I double checked and Vsync is set to "application defined" in the Intel drivers, yet it still syncronises itself, making me wonder what the hell Intel's drivers are looking for that it still thinks my application defines itself as being synced.

I tried this out with Minecraft and a few other games I had on hand and they all seem to behave this way as well.

However, Deus Ex GOTY, which uses Direct3D, doesn't.

While it's all but guaranteed to be a hardware/driver issue and therefore unsolvable, I'd rather give it a go considering how commonplace these drivers are. Question is, are there any other ways to disable Vsync in Opengl/Win32, other than wglSwapIntervalEXT(0)?

>>16856894

This is quite interesting.

I tried to intercept non-client mouse clicks (via WM_NCLBUTTONDOWN), delaying them until the mouse is actually moved (WM_NCMOUSEMOVE or WM_MOUSEMOVE) which works for a second or two, but then has the same issue. The window takes a second to realize it's grabbed but it does, even if you're not moving it.

It sends the following three messages at the same time as the window freezes: WM_SYSCOMMAND, WM_NCHITTEST and WM_GETTEXT.

The first one is the most interesting as it deals with mouse and window movement quite a bit (the second one deals with mouse positioning as well, giving more credence to this). It uses a wParam to determine what command is requested. The wParam sent is 61458, which corresponds to F012 in hex. If you check this resource, F012 is not one of the available options.

https://docs.microsoft.com/en-us/windows/win32/menurc/wm-syscommand

Now, according to these guys

>Ah, I remember now… $F012 is a "magic nnumber" (not documented) used to drag a window

<So, when you use Perform(WM_SYSCOMMAND,$f012,0); you're actually moving the window. $F012 is MOUSE_MOVE.

>$F012 is SC_MOVE with $0002 OR'ed with it. The 4 low-order bits ($0000-$000F) of WM_SYSCOMMAND are used by the OS internally. In this case, $0002 means the Drag bit is enabled.

https://www.experts-exchange.com/questions/11544059/what-is-f012.html

https://stackoverflow.com/questions/3976610/moving-a-caption-less-window-by-using-a-drag-area

I'm considering writing my own window movement code as that will probably solve this. As currently, I have to DefWindowProc WM_SYSCOMMAND, as if I don't do that, Window movement is impossible. Shouldn't be too hard as far as I'm concerned.

>>16857028

I tried catching WM_NCHITTEST but it didn't solve the issue. WM_SYSCOMMAND with 0xf012 has the same issue.

>I'm considering writing my own window movement code

I'm aware of one person who solved this issue in the same way. He managed to rip out pretty much all of the default windows behavior in his windows, but he had to reimplement a lot of it to make it work. In the end his window was just one big client area and he had to do the platform work needed to give it controls and everything. But that was the point…

So, maybe that's all there is too it, and we have to take the plunge. I suppose that works, I may go down that path at some point. Right now i've just accepted the current limitations of my method, because i'm more interesting in developing my engine code.

>>16857066

It's actually not that complex. Just intercept WM_SYSCOMMAND and if wParam is 0xf012 (61458) tell the engine that the window is being moved and don't let it do anything else. Otherwise, just refer it back to DefWindowProc.

Then, if the mouse is moved, move the window accordingly. I just get the mouse position from WM_SYSCOMMAND, get a new one when the mouse is moved, compare the two and move the window, then overwrite the old mouse coordinates.

Lastly, if the mouse is released, just let the engine know that the window isn't moving anymore.

My game doesn't allow resizing of the client by just stretching it (you'll do that via a settings menu), I assume that would require a bit more work but this was pretty simple. With this, I don't even need to spam the engine with WM_TIMERs as it never freezes either way.

>>16857089

Should note really quick, if you move your mouse too fast, the mouse escapes the client and movement gets glitchy. To fix this, you have to SetCapture(hwnd) in the WM_SYSCOMMAND callback and ReleaseCapture() in the WM_LBUTTONUP one.

>>16856330

and I am not so sure about the "art" part

>>16858522

>yandev on suicide watch

He's still earning more than minimum wage for doing fuckall, anon.

>>16858560

He's making 2 dollars more than a burger flipper in Commiefornia and at the rate he is going he will be making $10 an hour in 3 months.

>>16858522

I would be happy with that amount

>>16858580

>>16858522

>>16858560

his total earning are probably somewhere near 100k. But knowing him he's retarded and probably spent it all on shit he doesn't even need.

>>16858580

I checked recently and he makes $2,7k of free money since he doesn't put any effort.

>>16859282

Not trying to be unhelpful or rude, but knowing nothing about how the temperature interacts with anything in the game, I'd have to say that you'll need to decide that yourself. I don't really see how someone could give useful feedback otherwise. You'll certainly want how the game plays to influence the temperature curve and not the other way around, unless you're trying to simulate conditions in some actual climate.

I want to say "no it doesn't look too extreme" because I imagine you're going to have temperature effects for very hot and cold weather and want those to actually be in the game, but at the same time I don't see how one person saying that helps you very much, and it's certain you're going to have to balance in detail later on.

Are you concerned about lore reasons for having 95 degree heat in summer and slightly frosty conditions in winter? Browse wikipedia for a similar climate (looks like this "humid subtropical" climate region is close) and check the nearby geography against your game's area. There's a lot of real-world data out there that can help inform a decision like that.

>>16859282

looks fine to me

>>16859298

>>16859312

it just feels strange to me because temperature doesn't variate that much where i live (specially because we live by the sea, so it is more stable)

Temperature will influence the crops instead of the seasons doing so directly

In most games

Season -> Crops

In our game

Season -> Temperature -> Crops

Some crops will be viable both in spring and fall due to this, as long as the temperature match you can raise the crop, no need to wait until the calendar says 1st Spring to start raising spring crops, the death of crops from being out of season will also be gradual instead of everything drying up when the 28th ends

>>16855898

>The alt key locks my window and I never looked into it that deeply. Kind of annoying and I end up forgetting about this constantly.

That's not that hard to fix in particular, just intercept the ALT (VK_MENU) click. In most applications, they bring up the contextual menu so that's what I made it do. I also replaced the regular contextual menu by intercepting non-client right clicks in a particular way and the VK_APPS key. However, the contextual menu still freezes the window and I haven't been able to address that aside for WM_TIMER spam. Seems like none of the messages that are returned do much of anything when intercepted.

>>16859282

Any temperature can be written off as just "that's what it's like in the region".

Where I live, Summer and Winter are 10 degrees cooler and fall is 5 degrees cooler than your graph. From what I know from a friend of mine, the US west coast would be pretty close to your graph, maybe a few degrees cooler in the summer and five degrees warmer in the winter. So aside from the extremes being more extreme in your graph, it's pretty close to the West Coast. Might be even closer to the east coast from my limited knowledge of their weather patterns. Just consider what region yours is closest to in terms of the real world.

>>16859361

and if you make different regions having each their own temperature modifier, you could make it more dynamic, although it might be very confusing for the player if it's not properly explained or if it's hard to visualize

>>16859533

yeah, soil features and temperature will vary, this graph i posted is the default for the city area

>>16859565

holy shit dude

>>16859282

Are you going to include snow in your game?

>>16859632

yep, but during the peak of winter only, mostly 7 or so days of it, most of winter will have tillable soil

we want to make the city lakes freeze too during the peak.

>>16859565

and I thought this year was warm. Mean in Janurary may be -4 and -15 is completely normal but in 1925 we had 16 degrees C. GLOBAL WARMING IS TOTALLY REAL GUYSE

>>16859282

I'd say it depends on the setting. If you think that's extreme, where I live, there can be as much as a 90 degree (Fahrenheit) difference between the hottest summer day and coldest winter night. I assume your game isn't taking place in a desert though, but it's not like we don't have farms out here despite the climate either.

>>16860051

Objects themselves are fine, but the way they are commonly done (especially in C++) is idiotic.

>>16860051

using = for pointers should be fine, since all you're doing is copying a reference to your data, not the data itself. It should be very performant - you just gotta make sure you don't end up with memory leaks with your pointer use.

gay niggers lmao

>>16860051

That is the default temperature for each season, weather will increase/decrease the temperatures further

If there is one thing that our game won't lack is autism

I've decided I'm going to use my dungeon generation code as a base to make a 7 day roguelike. https://itch.io/jam/7drl-challenge-2020 Not sure if I'll be able to finish it in time but whatever. So since I have nothing to work with for monsters I'm going to one monster that will be the Minotaur. The Minotaur will hunt the player through the halls of the dungeon while the player tries to gather up as much loot as possible before dying. There will be just one decent sized floor with a few connected basements. If I can't get the Minotaur's navigation code working properly in time the backup plan is spectres and ghosts and stuff that can go through walls.

Do you think it would be more appropriate that the player have to get all of the loot and then leave the dungeon as opposed to just getting all of the loot to win?

>>16860488

Roguelikes usually have different tiers of wins. Getting out with 100% loot could, in your game, be a complete victory, whereas getting out with less would be less of a win. It'd introduce a resource management aspect by having the player decide when is the time to cut and run. Getting that balanced right in seven days, though, could require a lot more experience with design than I'm guessing most amateurs have, so it's probably better to aim for just getting 100% and getting out. Just actually getting all the loot and that's all feels a little simplistic, but should be fine if you're okay with players making decisions like "it's fine to dash right at the minotaur trying to scoop the last treasure before he beats me to death."

Having said that, with just one monster I feel like it'd be a good idea to get that one monster's behavior as threatening and intelligent as you can. It sounds like the game's pitch is basically You versus the Minotaur, and having that one adversary be worthy and intimidating will be what makes your roguelike stand out.

>>16860500

Yeah I'm thinking of having the Minotaur smash down walls between rooms and bring down roofs on top of your head and such while on the other hand the player can spike doors to keep them shut or drop ball bearings and such to get away. Also I'm thinking of having loot have different values so its up to the player if they want to collect all the lose change in a room while the Minotaur gets closer rather than leave it and look for better stuff.

Getting all of the loot will obviously mean you win but I'm kind of leaning towards players just trying to get a high score. Don't know yet.

>>16860514

Without a lot of really careful design and testing, you're more likely to have a playable game if you leave it open-ended and go for high scoring, I think.

>>16860488

>The Minotaur will hunt the player through the halls of the dungeon while the player tries to gather up as much loot as possible before dying.

Try to rip off Mr X or the nemesis from Resident Evil but make your minotaur smarter.

>>16860643

Never played. What do they do?

>>16860643

Big guys, who chase you around the level. If they hear a sound they come to you and try to kill you and you have to hide from them.

>>16860594

You only ever need at most one temporary variable to rotate fields in place, but in this particular case your rotation is an exchange, so you can abuse the fact that A = A ⊕ B ⊕ B, B = B ⊕ A ⊕ A. If X, Y have the values A, B respectively, then:

X := X ⊕ Y (now X = A ⊕ B)

Y := X ⊕ Y (now Y = A ⊕ B ⊕ B = A)

X := X ⊕ Y (now X = A ⊕ B ⊕ A = B)

will exchange the two. This can also easily be done in parallel for rows.

>>16861033

What I did before was have it swap with (y * i) mod n, so that each row is guaranteed to be visted and swapped. But then I saw that if the Y shift is a factor of the height (eg 2 and 4 on this case) then special cases come up

>>16861062

This happens not just for factors but for all shifts with GCD(shift, height) > 1. Four is just small enough for these two concepts to be the same here.

I think you can still salvage this algorithm if you correctly account for the fact that when GCD(shift, height) > 1, you have multiple disjoint rotations instead of one big rotation through all rows. Whether this is still efficient in the end is something I can't say without writing/seeing the algo myself, though. If you only want it for 4 rows, you're probably better off with a special case anyway.

>>16861103

Actually yes it can be saved, and it's not hard. The explanation can be simplified a lot with some basic group theory, but I don't know your experience so I kept it as elementary as possible.

Let s be your shift, n the number of rows, and d = GCD(s,n). Pick some row index i to start in and step through the row indices modulo n in steps of s, until you reach your starting index i again. This happens after at most n steps because i ≡ i + sn mod n. If you needed t steps, there are exactly t different indices in your cycle.

By construction t is the smallest number such that i ≡ i + st mod n; subtracting i gives the equivalent condition st ≡ 0 modulo n, which doesn't depend on your choice of i, so we see that all such cycles have the same size t.

Because you can step forwards and backwards, two cycles of row indices are equal as soon as they share a single index, and an index j is in the cycle starting at i if and only if j ≡ i + ks mod n for some k, or in other words if and only if j - i = ks + ln for some k and l. By Bézout's lemma, such k and l exist if and only if j - i is a multiple of GCD(s,n) = d, so the cycles starting at 0, 1, …, d-1 are distinct.

Summarizing, there are GCD(s,n) = d distinct cycles with starting points 0, 1, …, d-1. So the easy option is to calculate GCD(s,n) and do the necessary swapping. Another option is to abuse the following two things:

- t, the number of indices in a cycle, is also the number of swaps in a cycle plus one.

- The n indices split into d cycles of size t, so n = dt.

This yields the following code which doesn't even calculate GCD(s,n):

rotate(rows,shift) {
	const n := length(rows);
	var start, done, current, next;

	start := 0;
	done := 0;
	while (done < n) {
		current := start;
		next := (current + shift) % n;
		while (next ≠ start) {
			swap(rows[current], rows[next]);
			done := done + 1;
			current := next;
			next := (current + shift) % n;
		}
		done := done + 1;
		start := start + 1;
	}
}

Here we go folks, the bare minimum for a roguelike game.

https://mega.nz/#!9rQVDKha!H8SUE5Ct5HBKqEDTlWtzZWClRMNoorpyVVusCqM0A3M

If anyone wants to give some feedback that would be great. I still have until March 7th at 6pm to polish it and add more features. Oh also don't press the space bar, that just loads a testing scene.

>>16861247

It can be used with any dimensions so a general approach that isnt a brute force is helpful. I think that having multiple cycles was throwing me off.

This whole thing is to take linear array data, and interpret and manipulate it as rectangles of data (eg efficient tile storage in an array, but manipulate it intuitively as a map chunk)

>>16861632

Godot is weird. It exports the entire engine along with the game. The editor its self even runs in engine.

>>16861646

>The editor itself even runs in engine

That seems like a helluva gateway to modding games.

>>16861746

agdg is for modding as well as deving.

>>16861730

calling this music shit is a major understatement, how do you think this will attract any backers?

>>16861730

For a second i got excited, but then i saw the pause.

Also the actual models of the girls need to be cuter. They seemed very bland.

If someone could find me a good heavy footstep sound effect I would be very grateful.

>>16861978

shimapanzer. I think it never got anywhere. definitely more shit than UW

>>16861918

I hope the gameplay speaks for itself

What's stopping you from making a game?

>>16861730

According to some german "dota but" player, the kickstarter trailer is 95% of the success of the kickstarter, os they should've spent at least 2 weeks on that. If they did, well…

>>16850084

Bitch I work 55 hours a week. I don't live on daddies or mommies cash. Until them or you.

>>16850084

They go outside and make hipster friends who help them. Seriously. Bite the bullet and go to some hipster fucking shit like your local IDGA or Women who Code chapter or some repulsive shit like that. You will become a better dev for it if you can get in on some sort of group project or if it gives you a lead on a game jam.

Another possibility is they are genuinely motivated by their stupid ideas where you are not. Its one thing to try to pick up every nuance of code or blender or whatever and its another thing to just download an engine to sit down and just make something, even if it is shit. Your inner desire to not half-ass things is probably preventing you from even doing anything. Throw that aside and spend hours and hours on a shitpost in game form. You know you will enjoy it.

If you don't even know where to start, godot has great tutorials in its documentation, unity has billions of tutorials. Game maker as well. Also, RenPy if you want to do a visual novel which is arguably easier than making a website, but using the extensions people have made is a great way to learn python.

Basically: JUST DO SOMETHING

Quick feedback question. What UI options do you prefer in a Zack McCracken style adventure games.

- simple: look, use, talk

- complex : look, pull, push, open, close, lift etc.

I like the simple approach, as the complex one just has me trying out all options. Anyone has a different opinion?

I ended up remaking my game, for a demo, from scratch and I was able to recreate 1/5 of the game in 2-3 hours.

The main game took me 2-3 months to create.

I'm I just getting better at game making or am I just shit

>>16862527

You skip most of the involved learning and design effort when remaking something, so it's bound to be a lot faster. Just don't fall into the trap where you infinitely restart from scratch before your first release.

>>16862482

> as the complex one just has me trying out all options. Anyone has a different opinion?

how many objects do you have that would need different interactions instead of simply click -> "use" and do whatever makes sense? and if it's worth to keep the interface more cluttered for it?

>>16862482

Like the other anon said, it's entirely up to what you think you can do with it. If you think you can work in some puzzles where the difference between open and close and lift is relevant, then do it. In fact, if you do add the extra verbs, you'd better make sure they get used cleverly, but not so cleverly that you create terrible puzzles. Like, if you have one door that you have to LIFT instead of OPEN, unless it's because all other doors are powered and this one is depowered, or something, that's not a good puzzle. But maybe a door that you have to OPEN and then LIFT could be a decent one. You could also go a hybrid route, if your engine allows it. Like, have different classes of items give a list of verbs that could work with those items–like every door has both an OPEN and a LIFT option in a little radial menu when you click on it. That saves you from having to write a SMELL and a KICK response for every door in the world, but at the same time it allows for a little more interactivity than just one unified USE verb. Really, unless the nostalgia for many-verb interfaces is extremely important to your players, then it's entirely up to how much work you want to do and what sort of interactivity you think you can take advantage of.

>>16864975

That second image is giving me Xenogears flashbacks with that round, jaggy-bordered 2D sprite wandering around on a big square.

>>16865007

My game is based on Xenogears but with stealth and targeting like Fallout 1.

>>16865054

FUCK YOU I WON'T DO WHAT YOU TELL ME

>>16865054

>NOBODY BUYS GAMES FROM STEAM ANYMORE

Someone tell Valve that so they start making games again instead of just sitting on their massive money pile all day.

>>16865054

>YOU MUST REINVENT THE WHEEL BECAUSE REASONS

Autism reasons.

>>16865071

MOTHER FUCKEEEEEEEEEEEEER

>>16865054

>DO NOT CODE W C++

>CODE WITH C

Why? I totally understand not using an engine as those can really limit you, but why bother going with the less-feature-packed language? So you could waste time reimplementing the features on your own?

>>16865157

Just do it all in Python bro. No need to reinvent the wheel.

>>16865161

Python's indentation makes it unusable for me.

>>16865145

Memes aside "reinventing the wheel" implies that there is something as perfect as a wheel, but Unity and friends couldn't be further from perfect. The expression also implies there's nothing more specialized to invent, just because truck wheels exist doesn't mean you shouldn't invent a new wheel when you want to make a bicycle.

>>16865198

Creating an engine to make a game.

Writing an image editor to draw a meme.

Writing a DAW to write a song.

All of these are basically the same thing. There is nothing inherently wrong with writing an engine to make a game. It is just completely unnecessary and if your goal is just to make a game it is a complete waste of time and effort.

>>16865173

>the only reason

There's a lot of other reasons to enginedev. You have more control over each part of the system and it might be easier to do changes to things because you understand it thoroughly. You get a lot more programming/dev experience (and the experience of learning and finding solutions itself) if you build the whole engine instead of just doing high level unity stuff. You won't be bound to a third party in case they cuck out or the engine goes to shit or becomes paid-only or adds spyware into your games or something. You won't be limited by the platforms that the engine supports. Your game size will be significantly smaller no matter what kind of game you make. The performance is very likely to be better all around even if you're not an optimization freak, just because the engine is tailored to the job it's meant to do instead of being this generalized do-it-all abomination that was bent to do it. If in a distant future you want to do something that a premade engine gets in the way of, you're already equipped to do it. If you find that developing tools or other programs is more interesting or profitable or just helpful for gamedev, you already have basically all the skills to do that since your engine is just another form of it.

Acting like doing it is a waste of time sounds like butthurt that some people can do something harder and "ideologically" better than you can. You should ask yourself what you want to do. If you just want to shit out games easily or want HD™ Realistic graphics and don't care about anything else, then there's little reason to enginedev. Enginedev is more for investing in your broader long term skills as a programmer and your ability to do high quality software, but isn't practical if you're a brainlet or don't have some extra years to spend before you can start making games.

>>16865198

Keyreal is working on his third game made using his own engine, with most people here knowing his second one (Speebot). Part of the reason he can finish games as fast as he does is the experience from knowing how to mke an engine, plus the fact that his engine works the way he wants it to.

Difference between Keyreal and for instance Sigmadev is that one finished their engine to make games, and the other works on his engine to make the perfect engine for what he wants, which is targeting nearly every PC from the 90s until now.

>>16865218

>You have more control over each part of the system and it might be easier to do changes to things because you understand it thoroughly.

In theory, but in practice you can usually work around it. For example, I didn't think stealth was possible in the Source engine due to how the lighting worked until I saw Dark Messiah do it.

> You get a lot more programming/dev experience (and the experience of learning and finding solutions itself) if you build the whole engine instead of just doing high level unity stuff.

This kinda falls under autism for me. It's really debatable how much of that experience is meaningful. You can get problem solving experience multiple ways.

> You won't be bound to a third party in case they cuck out or the engine goes to shit or becomes paid-only or adds spyware into your games or something.

> You won't be limited by the platforms that the engine supports.

There's tons of open source engines.

>The performance is very likely to be better all around even if you're not an optimization freak, just because the engine is tailored to the job it's meant to do instead of being this generalized do-it-all abomination that was bent to do it.

True but I don't really see a lot of people here doing shit that is such a performance hog it requires it's own engine. Unless you want to make a game that runs on absolute bare minimum hardware which falls under autism for me

> If in a distant future you want to do something that a premade engine gets in the way of, you're already equipped to do it.

You can do it with an open sourced engine to, but like I said you can do a LOT with engines if you're clever enough. Not really sure if there's any game today that warrants it's own engine aside from MMOs. A lot of old games are being ported to unity.

>If you find that developing tools or other programs is more interesting or profitable or just helpful for gamedev, you already have basically all the skills to do that since your engine is just another form of it.

Breaking into that field is not easy. Recruiters don't give a shit about projects they only care about real work experience and if you're doing it solo you have to make the connections.

>Acting like doing it is a waste of time sounds like butthurt that some people can do something harder and "ideologically" better than you can. You should ask yourself what you want to do. If you just want to shit out games easily or want HD™ Realistic graphics and don't care about anything else, then there's little reason to enginedev. Enginedev is more for investing in your broader long term skills as a programmer and your ability to do high quality software, but isn't practical if you're a brainlet or don't have some extra years to spend before you can start making games.

FYI I'm an engine-fag porting my engine from C++ to C.

>>16865239

I have no idea how that dude does it. I know he works a software job, but most ones I've seen are so soul crushing I never have the energy to do anything afterwards. I don't even play games that much lately.

>>16865218

>You won't be bound to a third party in case they cuck out or the engine goes to shit or becomes paid-only or adds spyware into your games or something.

pretty much irrelevant, especially for engines that are used on a corporate level. the minute they force changes retroactively, even if it just license, is the minute they lose all their user base. no company is retarded enough trying to fuck with customer's production line, so at worst it's upcoming releases, which means you can switch after the game is released or your project's done.

like most shit in the end it depends on the scope of what you wanna do and what you need for that.

>>16865239

Did he use OpenGL or did he design his own renderer? Did he program his own libraries? Did he write his own compiler? What about the language he used? Did he design his own to efficiently describe the game he wants to make? What about an OS? Did he create the perfect OS on which his game can be ran on? Honestly though, he can't start making a game before he fully understands how to create a computer from only logic gates, and even then he should have a firm grasp of how to physically distribute transistors to create his own computer architecture. And even then you should fully understand the process of making a silicon chip, because what good is a chip if you don't know what is in it?

Knowledge of lower level systems is going to help you as a programmer but for most just downloading an engine and getting started on simply creating assets and writing systems is a great starting point for anyone who just wants to make a game.

fuck me, anyone ever done pathfinding on a grid? different movement costs, so using dijkstra. but the grid itself is changing too and I can't wrap my head around how to do dynamically calculate the graph when the grid gets bigger.

>>16865801

a* pathfinding, I think it's slower than dijkstra but provides a good short path https://en.wikipedia.org/wiki/A*_search_algorithm

boundry box if possible

multithreading

if you're moving multiple things in real time, it gets exponentially harder.

I added a new thing in the wiki for enginedev:

http://8agdg.wikidot.com/programming#toc6

Feel free to add your hot opinions about why it's good or bad.

>>16861746

Nice, some screenshot would be good to see.

>>16865054

Use what works, draw off of references if it helps and if it doesn't work or requires ten years of work to make feasible stop and do the easy option.

If you code, learn better formats. If you do 3D models or animation, use blender and don't shell out for an expensive animation software, it also does 2D animation now. Blender isn't hard to use, that's a meme likely perpetuated by software companies to get you to buy their shit.

Avoid excessive bloom and learn to optimize your shit.

>>16866482

Oh better a video, I hope you don't stuff it with too many hitscanners enemy as it would be a annoying trying to dodge on those narrow area. Reminds me of a few russian mapsets where the player spawns in a apartment.

>>16866492

Binary is already a language. Only by using an analog computer can you truly make a video game.

>>16865809

a* is a bit more complicated and wouldn't make that much of a difference considering the grid size I'm aiming for (plus some wonky calculation with movement costs I'd have to work around as well).

still need to figure out how to do the graph anyway, most stuff I've looked at simply builds it via array and from there the visuals, while I'd need to do it the other way around due to a non-uniform and growing grid. was thinking of simply collecting each tile's coordinate into an array (which would have the bonus I could easily set the padding between tiles), but haven't tested yet how slow and what a performance hit it might be building the graph this way.

>>16866368

https://www.doc.ic.ac.uk/~susan/475/unmain.html

>>16866465

>Feeling like you *have to* build the engine is an outdated mentality.

imho it simply comes down to what you want to do, and that seems to be the difficult part for most people.

if you just want to make a game and see if your ideas work out, it's simply overkill. if you're into the technical aspect, why not, but of course you won't make much for a while (of that is your goal to begin with).

>>16866558

>Blender isn't hard to use, that's a meme likely perpetuated by software companies to get you to buy their shit.

usability is hardly a criteria why companies use several expensive programs at once.

but yeah, if you're a solo dev and just starting out blender is more than fine, and once you learn the principles it's not hard to apply them somewhere else.

>>16868073

Oh also there a few dungeon gen bugs that rear their ugly heads every 10 or so runs. If you are doing super good and havent seen the minotaur he's probably walled in. Or you are walled in, however you want to think of it. Just restart and it should be good.

>>16868073

Who says you have to stop working on it? Declare this version 1.0 and say that only 1.0 is submitted for the 7DRL contest, then just continue on making improvements. I don't think you should feel you have to abandon a project just because it's being submitted to some contest. Sure, you'll have to resist the temptation to resubmit new versions, but that's all on you. You'll always have version 1.0 to view as the thing that you were able to do in a week.

>>16868108

Yeah I just meant that I can't submit any changes to the jam. I'm going to keep working on fprl. The Minotaur's Lair was just a fun experiment.

>>16853633

fuck, that looks beautiful

>>16853505

wait people have made small stuff in months, are your guys looking at how the devs that came before you did it?

>>16865054

C is just C++ with less tools. Most C code compiles fine in C++ with minimal changes, even shit from the 90s. I should know, I worked through Petzold's book last year and all his examples work just as well when compiled in C++, with maybe an explicit type cast needed to be added in to a few lines here and there.

Even if you only use a few features from C++, it's still way better than C. You can write C like code in C++ and take advantage of C++'s superior libraries streamlined features like range for loop.

>>16868476

I, personally, go over it with a few steps. First, I lay out the general summary of what I want my game to include. This means things like the the overarching goals and tidbits that make my game unique, not the precise mechanics. That comes later.

After that, I lay out, in sections, each mechanic to be created. I usually do this in an alpha format (so section one is alpha v 0.1), but you can do it however you like. I then give that alpha a "theme" as it were, and fill it with the appropriate designs that match that theme. You then simply go down the line, creating sections and designs until you are satisfied with what you want to put in. Take a look over what you have, cut down any unnecessary fluff or things you believe outside of your skills for a reasonable time frame. Then, stick to it. Modify if you have to, depending on you growing design philosophy, but don't add anything. If it's absolutely necessary, like you forgot something, then sure. But if it's just "cool" or "I'd like it to be in", then don't. You have a goal for your game, and what your game should be, and you need to reach that goal without useless distractions.

This is more a design document strategy than a design bible, but I think it works just as well.

>>16868895

Sex mods when?

>>16868912

Screen fades to black when the Minotaur gets you. Use your imagination.

Some more footage of my catgirl platformer.

Added sound effects and an NPC character. She'll be used for testing out dialogue boxes later.

Kasha's run animation also needs to be updated as it's a little stiff. I already have a newer one where her head bobs a bit more.

>>16868351

The reason you'd want to use C over C++ is to avoid the libraries and features. They're too over generalized to save CPU cache/System calls, which are a large bottleneck in programming these days. There are objective reasons to use C over C++ it's just they're very niche and only if you want the absolute highest bleeding edge performance.

>>16850084

SUBSIDY

>>16869482

What next, ban all questions because everyone in here is too much of a n00b to give valid answers?

When I was starting out I had a lot of questions that I couldn't find answers for. Someone who's just starting out has very little to no understanding of game development and even less context for how useful it really is to make your own engine. You might look at Factorio moving millions of items per frame on toasters, and then compare it to a bloated Unity game that requires high end PC horsepower just to walk in an environment, and then think that engines are complete shit and go on thinking that you need to make your own engine to have any kind of technically sound game.

You only need to be above the "starting out" level to be able to give at least some kind of advice to someone who is at that level, besides some of the advice might have been adopted from other much more experienced developers such as Casey or Jonathan Blow who stream often or C++ DOD man whose name I forgot. All 3 of the aforementioned have decades of development experience, have worked on highly successful and/or AAA-tier games, advocate for either C or C++ without most of it's features, and think that making your own engine is valuable.

Usually the worst advice is to tell people to do this and don't do that. Instead you should weigh the advantages and disadvantages of things and let each person make the decisions on their own.

>>16869663

>What next, ban all questions because everyone in here is too much of a n00b to give valid answers?

You're missing the point- that anyone capable of making a game is also capable of answering this question- thus any debate about whether being an enginefag or not will or wont result in a game is irrelevant because the person listening to the debate is not experienced enough to finish their projects.

And are these really valid answers? It's just people butting heads over what skillset they decided to specialize in, trying to get everyone else to use "their lane" because they understand it and don't know why people would do something else. The big argument above us really is just a "do this and don't do that" argument. A "you care about X but I don't care about X so your point is invalid…", etc. All presented as actual advice or… answers. Stuff that can never be useful because its so decontextualized from any real scenario that it doesn't matter.

>>16869482

You're right about it as time is everything in game dev. You're not wasting your time if you're using an engine like Unity/Unreal as you're exposed to the latest tech while learning a industry standard workflow that can be adapted anywhere else. It'll even help with enginedev once you encounter limits. A lot of great programmers working on complex projects, started from simple ones or fun as it motivates them to continue programming.

The excuses of "bad optimization" is a cope beacuse you will improve performance from feedback by consumers. Especially if it's a well documented engine like unity/Unreal.

>>16869862

Purpose of Unity and "industry standard" workflow is for managers to minimize risk within a project. If you accept the speed loss, and follow the guiderails, you will be slower (inherently) , but optimization is about meeting your performance target, not being "fast" according to some arbitrary metric. You should be able to define how fast or slow a project will be in-engine for 95% of the kinds of games that can be made. If you cant predict how the engine will handle the game, that is a good sign you might need to use your own engine.

Latest tech is almost by definition something that is more advanced than what exists in Unity/Unreal. The industry standard workflow is in fact the limit on technology, has less and less to do with the state of graphics programming as hardware gets faster. Most innovative techniques do not scale to actual games because the toolchain for them does not exist.

How difficult is it to make a VR sex game?

>>16863464

>>16863335

Thanks for the feedback. I think I go with a simple UI. Look, Use, Talk, Take and perhaps a Combine command for Inventory. The puzzle needs to be in the items and scenes and not in the UI.

>>16869945

Not difficult, as you get automatic sponsorship from Tel Aviv University.

>>16870750

Nice "I didn't read the thread" post retard . I'll give you some help, the discussion my post is a response too starts here: >>16865054

Only takes a few minutes to avoid wasting your time on a (1) and done response that has nothing to do with the topic at hand. Not that the post didn't already establish itself as a contextualized response to an existing debate and not some vauge platitude about how nobody here is qualified to discuss game development.

>>16870756

>>16870750

Also, because the thing you quoted might be vague: people are capable of giving advice on problems that they've already solved. 95% of the people here have not solved the problem of "using a tool to make a game". The ones that have know better than to give decontextualized takes about whether coding everything in C or using Unity because writing engines is "outdated". It's just useless advice to give without any context behind it; people reaffirming the value in the skillsets and ideas about gamedev they've invested themselves into. The blind leading the blind …

>>16870756

> I'll give you some help, the discussion my post is a response too starts here: >>16865054

Thanks for reminding us how autistic you are.

Everyone is wrong, including me.

>>16871462

IMO making something look good is a bit harder in 3D due to the animation part and how you play it, while you can "just arrange" it in 2D with different frames; but both approach is hard regardless.

Maybe adding some background layer would help, as in having "far" background and nearer background object.

On my end, anon shmup reporting:

Code at work didn't need much fiddling, but didn't feel well for february inb4 corona but felling better now.

At least fixed a bit trails management, it was working somehow by luck, but now it's better, still with some glitch, probably when looping around.

Also worked on attaching trails/particle generators on models, having trails that change shape, and currently working on using texture on trails, it's almost there but the dynamic UV texturing is still wonky.

Also arranged a bit materials management, before it was 1 material<=>1 possible texture, but all textures don't need to be "high" res so it's wasteful. Now I can have multiple material reference different part of a same texture, and I could after try different texture for my explosion since it looked interesting when I was debugging some test trail.

>>16871462

I think there's a lot of ways to go about it but me, as a programmer, mostly went into the technical details of art. Setting up a palette, resolution, pixel densities etc. I think that's a good approach if you're not very artistic like myself.

Any opinion on input libraries for C++? I'm used to both SDL and SFML but I just need input.

>>16872236

Can't help you much, besides saying that I've used SFML with C# and enjoy it.

>>16872236

https://github.com/wgois/OIS or https://github.com/jkuhlmann/gainput

I personlly like the latter, Gainput, but it's not actively being maintained so as time passes it might become less useful. But at least it doesn't shove a COC down your throat.

I started to learn how to make shit 3d models.

I don't even want to think about 3d dev right now since I have to use a new (aka a 'real' engine) engine that supports it and all the faggot shit that comes with it.

>>16872348

Lasagna.

>>16872348

dude, weed, lmao

>>16871714

It's hard to quantify the hardness of those two different approaches, but yeah it seems quite similar. I better stick with what I know best.

Also, RIP.

>>16872236

I never understood that "palette" thing. It makes no sense to me.

>>16872502

>I never understood that "palette" thing. It makes no sense to me.

It's just to limit what shades you use to make it more cohesive. It's also about keeping even quality over time and multiple people. You sacrifice artistic freedom for better productivity since you can't really question what colors to use.

>>16872502

Working with a limited palette is useful, even from a strictly art perspective. Grab maybe 5-10 colors, and you can blend them in varying amounts. Think of them as your base paint colors you squirt out a tube, as if you were an artist. Like anon said >>16872541 it helps you maintain a bit of cohesion even if you blend and alter many variants of that palette together.

If you're making a game with pixel art, a stricter palette is essential for each thing you're making. Outside of pixel art, it's a helpful guideline.

>>16873122

Don't mistake a palette for a style guide either. The style guide would include something like a few color samples (that you can use as a base to build a palette) and say shit like "Outdoor scenes in Havenwood should be gray and brown and look dreary to reflect the miserable mood of the inhabitants. Even highlights and important objects should be cool-colored to emphasize this (such as the roof)"

Finally ported my shitty maze game to windows. Time to work on a real game and stop doing autistic engine dev.

Has any enginedev tried Direct3D? How is it compared to OpenGL?