/AGDG/ Kowloon Walled City Edition

Might and Magic had the Learning skill and it was amusingly useless, despite giving up to like 30% more xp

Plastic butter knife
Its dangerous and lethal, thats why its regulated in real life

8ch.net/v/res/12241548.html

You think it'd be profitable to make a proper Valkyria Chronicles clone but featuring nothing but Mega Milks you can bang? This is same user who was thinking of making one since 2-3 years ago.

But THIS TIME, it might actually be possible, since upcoming UE4 4.16 will have fucking built-in rigging and cloth tools. Which means all I'll need to do is get a generic waifu model, autorig it, then paint cloth on her.

I've actually tried making some VC-like combat style before, it is doable, if a bit clunky. But the feel is there.

I think this is so the only way to honor and do justice to the countless Mega Milks butchered by Sega of Japan.

Fam

Also sorry for the Reddit spacing. This is just the way I type, I don't use Reddit that much.

Considering that the player will be deflecting bullets with this, a plastic butter knife would be pretty funny.

3:

You're approaching this from the wrong angle.

Reminds me of this tip Sid Meier has shared in some game design talks regarding balancing. When it comes to adjusting values, "if something is to strong, half it. If something is to weak, double it." I think it is decent advice because it is easy to start fiddling with values making changes that are to small to have a meaningful impact and wastes time.

So basically, binary search.

Well, I quite like the idea of how Dot Hack handled XP, in that the amount of XP needed to advance in level was always 1000, but the amount of XP received from enemies was determined by the difference in your levels (i.e. same level as you, get base xp; higher level than you, get more xp; lower level, get considerably less xp). I'm using that system, plus I'm the same user from , and am having less quantities of enemies to fight at a time. Also, it's not that INT only affects XP gain, it's also needed as pre-requisite for investing many feats that open up new gameplay mechanics; but if it makes any difference, the XP gain is currently set to 2.5% per modifier. The modifier formula is the same as the one for most table tops, [(Attribute / 2) - 5].

Also, I should mention, I don't know yet how well this will actually play out, but my original intention is that on Leveling Up the player gets one point that they can either invest further into an Attribute, or unlock a feat/perk/spec/whatever; point redistribution is not allowed beyond Level 1. At Level 1, all attributes are preset to 10 (+0 modifier) with no feats, so redistributing those points will impart negatives in at least one area, with different consequences for making any one of them a dump stat.

So far I have a basic leveling system and UI in place, but still need to figure out how best to make Perception have a positive/negative effect on the player based on their modifier. I want to make it affect the player's sight/hearing, but I'm not sure how much of that can be modified on an individual (Player) level and still be fun. And Strength, but I guess I could just make that modify recoil in someway; I kinda like the idea of a gun knocking you on your ass if you have shit Strength.

lets hope for that one user to stay strong

Man wish I was aware of this method few months back. I ended up filching the XP table from Pokemon for my game.

I thought Pokemon did that too, though. Or do higher level pokemon just give more XP outright, rather than relative to your level, so it seems like it's relative when you kill them with a low-level pokemon?

It's a really nice system, and I'm surprised I haven't seen it used more often. Too bad the quadrilogy its originally from is really grindy; although the gameplay in the latter trilogy series did improve things considerably.

The page this is from isn't up to date, but it should help.

So will enemies scale in difficulty based on their level? I assume not. Either way, what that means is you're punishing players for getting ahead. Giving more xp for overleveled kills is nice, it rewards skill and allows you to soldier through difficult opposition, rewarding you by helping you catch up. The upper half of penalizing xp gain against underleveled enemies is bad. Look at it this way. Assume an area contains a certain set of enemies. You can kill either all or the bare minimum to progress, possibly none. If scaling is a thing the latter is preferable since it's faster. If not the former is nice since it allows you to overlevel, investing time now that saves you time in the future. By penalizing overleveled characters you remove the reward of clearing the level and sidequesting (unless these activities reward in some other way, ex. giving you good gear, upgrades ec). But if so, why the arbitrary split between gear and levels? Why not reward upfront with xp?

Herp. What I mean is "Will enemy difficulty/level dynamically change based on player level, possibly leveling up when the player does?"

if your referring to levelscaling which was the cancer that made Oblivion really annoying, I doubt you'll find anyone on this board arguing for it. And removing the xp gain for overleveling is as old as Diablo 1, but the reward for clearing levels was still there in form of loot.

but what if i wanted the second girl.
first girl has cowtits and tumblr hair.

How goes the floppy semi-erect estrogenised tranny penis physics /AGDG/? Remember, it's Tranny Tuedsay, slap a trap, or fap a trap, it's all about the feminine penis today!

i can't decide whether to have a tranny in my game or a trap. it'd be redundant to have both.
see, with the trap, , his trapness is going to be a secret, and he's also gonna be a prostitute, and go to the same school as you. so you can actually tell your classmates about him being such, which then drives him to suicide.
i haven't thought of anything interesting for the tranny though, so i'll probably just use the trap.
because the society in game is degenerate. it's the same reason i need an illegal in there as well.

open your dope sheet, select your frames then drag them to the time you want

JUST

The Trap can just be a tranny user. After all, Hormone Blockers at a young age don't make much difference to a crossdresser, they won't have tits by that point.

he's 16, user.

16 is legal for raping in the UK tho

There's no legal age in UK depending on where you come from

Rape is just legal, period, in the UK, at least if you're non-white.
But this isn't the UK.

says you

i know the setting of my game better than you do, user.

Get the fuck out. You fucking faggots manage to derail every close-knit thread with your shitty fetish. You forcing your awful dick girls into everything is no different from forced diversity or disgusting furried.

user, at what point did i say that i like traps.
me hating them is the reason they're in the game.

This.

Dickgirls and traps have nothing to do with gamedev and only serve to push the thread closer to the bump limit. Fuck off and reported.

What gets me the most is that fags like aren't even here to talk vidya making. This person isn't making a fucking game they're just here to spam their cuck-tier fetish. It is not only disgusting and horribly halfchan-esque but it actively drowns out nodevs and spits on our work and image.

You're still bringing in cancer and responding to someone who is not even making a fucking game. AGDG is one of the only good general threads to exist and we will not let slack jawed faggots ruin them.>>12246375

but longer barrels give bullets higher velocity as they have longer to accelerate up to speed. a sawed off shot gun with a 13 in barrel compered to the standard 26 in the difference between loads rounds out to being about 15% of the velocity

use social skills as dump stat strong and agile and sneaky as a cat, cannot check out of store without dropping the spaghetti. no one gives you jobs no quests just stealing food cat bungling grocery stores for food. you could go loot the cash registers but no one will sell you stuff. the more you steal from a store the more security they use. if you want meat, you need to break into the freezer in the back that has a lock camera door key all needing to be done/found before the first shift starts. cant even strong arm robbery because the very presence of normal fags gives debuff.

My first try with a RPG Maker

Alright, so I could make it so modifying the weapon's barrel would affect range, damage falloff, and impart a percentage damage modifier. However, the base damage stat would still be derived from the ammunition used, and a player's own stats still won't affect damage output.

They have their own combinations of Attributes, just like the Player does. I'm planning on giving the AI more challenging, unique behavior as you level up, but most of the stuff regarding that are just ideas at this point; I'm currently just getting basic stuff up and running right now.

I'm not quite sure exactly how the game is going to be setup yet, but I don't think having an open-world with level-scaling enemy encounters would be the way to go. If anything, it'll probably be closer (again) to Dot Hack in that you have a Nexus area from which to buy/sell/craft shit, and then warp to a scenario of whatever type and level base the player so chooses.

I know it hated like poison; and with all the people using default assets, quite rightly so, but how is it to use? Is it hard to add new mechanics that the engine doesn't cough out by default? Like having individual pieces of gear gain EXP.

Even colored squares have more personality than those.

You can make scripts in ruby if I remember right, if someone hasn't already implemented exactly what you're looking for.

and heavy barrels affect bullet spread giving a slight modification to damage based off of the weapon shooting it is not wholly unreasonable especially when you compare some stub nosed revolver to a rifle shooting the same round. having a standard and moving from that based off of barrel length sounds needlessly complicated especially when balancing but under your suggestion people would never carry rifles, ww1 would have been fought with a bunch of infantry toting pistols as the standard issue

You misunderstand me, what I intend is that there are three particular stat categories: Player, Weapon, Ammo. Player stats affect just the player, what they can do, and how well they handle weapons, but have no effect on damage output. Ammo stats have a base damage, range, spread, etc. Weapons have their own base stats (weight, fire rate, reload speed, precision, etc.), but also have modifiers on some of the ammo's stats.

The point is to be able to modify weapons and ammo to some degree.

Neat. I'm softened up using it.
Still can't do jack with it without someone else making decent sprites/graphics/music/etc.

How easy is it to get someone to make shit like that for you if you pump out a demo with default assets?

hmm alright so for example I have Run1 and Idle2, so Run1 has a frame length of 8 and Idle2 has a frame length of 20, so when I set the render time to 20 and setting back to the Run1 animation which has only 8 frames blender will render the 12 frames of the same 8th frame anyway, and I meant to ask if there is a way to automatically change the frames to render based on animation action length instead of a global one.

...

So what are the levels going to be like? Flipping switches to enable/disable certain sets of enemies might make navigating traps and hazards interesting.

Thinking about how to manage my shoppers needs and AI…

Components/ECS seem like the way to go, especially when it comes to saving and loading; not worrying about constructing complex objects and just feeding data into a system constructor seems easiest.

Like if I had a "needs" meter sort of like the sims, for hunger, thirst, sleep/energy, etc, then these could just be a bunch of bytes accessed by an entity int index. Also affects AI, eg more willing to pay for a burger when hungry or whatever.

Another thing I considered, people at work tend to ignore the top and bottom shelves and focus on the middle. What I could do is use their height as the midpoint of a normal distribution, and use that to determine where they are looking while going down the aisle.

Looking and product impression calculations would only happen once per tile moved (or better, once per entity turn - speed could be affected by mood or hunger etc and makes it so they move more units or tiles per tick)

there's good papers and info on how the devs did the sim's AI.
It's based on the utility AI approach, with a mix of a "smell" system for each of their needs.

Probably would match quite nicely with your planned game mechanics.

god fucking damn, why did i have to find a job, my dev cycle has been slowed down to a crawl

That sounds really intricate and encourages tedious micro-management. Putting different products on different shelves doesn't really sound like an enjoyable mechanic, even if it were implemented well. Shelves generally always have the same product, and a customer looking for a specific product will be OK with reaching up/down to get them.

Same but with uni. Last time I worked on it I got an unexpected segfault, and now I have too little time to figure out where the problem was.

Not sure what you mean by this. That item location doesn't change? If so that's flat untrue.
Sure, IF they're looking for a very specific product. He's not wrong that it makes a huge difference.

Youll be able to set 4ft sections, the shelves per modular unit, and the products on those shelves. Think of it as a templated version of DF stockpiles but with multiple items.

If I keep it to 4 ft sections then I can do theming bonuses on a mod - setting 50% (maybe 65% or 80%) of it with a common produce theme gives a bonus to product perception, making it more desireable. Eg everything is by one brand or has a common flavor

Take for example a shelf with soup on it. There's going to be cans of soup in the middle shelf, as well as on the bottom and the top. The whole shelf will be dedicated to one genre of product. Sure, the store may change where the soup is located from time to time, but the soup will take up the whole shelf every time.
Sure, for some really niche product, the stock might only take up one portion of one shelf, but that's an exception rather than the norm.

That's not consistently true at all. It's a nice thought but space limitations vs differing sales and margin expectations, vs the question of how many varieties of a particular category you might need to keep regular customers happy make it a constant compromise.

Are you planning on dealing with other departments at all?

i'd say it depends on the department.
just thinking of how our retail store works, the outdoor sections would fit your description. one shelf for tents, for sleeping bags, for mattresses and so on… but the sports department, it's very much portioned like you described.
and if we're talking about food, then think about how a typical milk aisle looks. butter, mayo, curd, sweetcream and sourcream all in the same general area.

It sounds like accuracy to me. Products at eye level make the most profit, as well as products at the ends of aisles.

I suppose. But from a gameplay perspective, it seems easier to just place one product on one shelf, if you're going to be having 400 shelves in your store. Focusing on the color of the package and the angle at which cereal mascots look, like in is just overwhelming when you scale it up to a full store.

It won't necessarily be tedious; just a little something to keep in the back of your mind when you want to minmax a certain product. It would also make objects like stepstools have a use in game, and you could avoid it by preferring taller employees. Of course, customers might need help, too. But we'll see how much of it gets implemented.

At the very least, I want some weighting towards the middle.

I want to keep it to foods to start, and then expand the game in arcs, like cosmetics, pharmacy, clothing, etc.

Oh yes, and I'll also have a tagging system for foods, to define their attributes. Think how boorus work, and how tags can implicate other tags, and you'll get the idea.

Thus, I tag one item as "organic", and it would also be considered "food". And some customers, based on their personalities would value (or not care) that it is organic, fair trade, locally sourced, halal, kosher, vegan friendly, and doesn't cast a shadow. But any "values" they have, they would much more likely buy things that match.

If you are doing tagging, maybe instead of such a large flat bonus to a whole section, having stacking bonuses based on how many attributes (especially visual) they have in common with adjacent items. That way it wouldn't be such an all or nothing decision for the player, and might give them a chance to get creative with their displays. Also, it wouldn't be that computationally expensive to crawl a 4ft item by item since you'd only need to do it when the merchandising changes.

This is why so many people don't make it.

I will find you, Jim

iktf user. i've been staring at the same screen for days, drawing fucking dialogue trees. and i'll need hundreds of these.
user help.

so i'm not the only one having that?

Anyone got any info on writing a multiplayer server?

I'm writing one atm now with C++ and PocoNetwork lib but I'm doing a lot of guess work. Seems like a lot of game networking is secret sauce that isn't too readily available.

Nope. I'm pretty sure that both issues have existed since Hotwheels was in charge, actually.

The amount of fun I had modifying this shitty copypasta was quickly overshadowed by the amount of work cobbling together the bloody remains of three different menu systems to form something workable and how half of the features in it don't even work as I wanted and another half of the features I want for it I can't think of a way to decently put in.
Fuck me.

Here, here. This is a large part of what I've been dealing with lately, and it was absolutely infuriating trying to find guides that are well-written, let alone ones that with the kind of game setup I needed. I'm told that the method GML uses is the same conceptually as C++, so I may be able to answer some questions.

rig is insisting that the spine rotates rather than bend. using the same sort of ik set up as the tail its seems insistent on fucking itself rather than just bending it says fuck it ima guna rotate like a bitch

just fuck my shit up

What about bone constraints?

What's the basic architecture setup of your server?

Atm I have a client list and two extra threads running receive and send calls.

the receive thread takes new clients and adds them to the list, if they are already exist it updates the clients data and resets their timer. the send function sends the data of all clients to each client.

the main thread loops and keeps track of time updating each clients current timer, after 5 seconds of no receive a client is determined to be disconnected and removed from the client list.

that's what I have so far.

I always forget whether SDL or SFML is the recommended thing.

Also, as a companion/standalone language for light games, is Python or Lua better?

>about to add one last missing feature and ship the plugin
I've actually finished something for once, even if it's just a 71KB engine extension.

if you want to make 2d I recommend Love2D.

I haven't tried pygame but it's probably ok too.

SFML/SDL are a lot of fucking around to create basically an engine that Love2D provides for you.

Seconding Love2d if you want a simple game framework. I like SFML and found it easier than SDL to work with on its own, personally. SDL/SFML will handle things like window management, drawing to the screen, playing audio, and some more stuff, but none of it is specifically intended for games though easily extendable to them. Love adds another layer of abstraction (it's a wrapper around SDL) that makes it a lot faster to make games with in a semi-standardized fashion with a lot less code. Lua/LuaJIT is better than Python for games.

magnificent ID

This is what I currently have, but it will become more complex than this, for reasons I'll explain at the end. In my game, players host, but the basic idea should be pretty much the same if you were going to have dedicated servers. As far as I know, the only difference would be how you handle the first player connection, which in a game like mine, would be guaranteed to be yourself.

Connection:

The server is the only place where the game logic actually runs, and the clients are simply mimicking the server's game state. Player inputs can be thought of as requests to the server to change the game state, which the server then decides on whether or not that request could, or should, be honored. If you're the sort who likes analogy, a way to think of it is that the server is a movie director, and all of the clients are actors, who do their best to follow his instructions and create the movie he imagines. Player inputs are like suggestions on line readings, which the director then takes or leaves. The director's say is final, when he finishes deciding what it thinks about that suggestion, it tells all of the actors exactly what to do. Generally speaking, it goes like this:

Player Input:

This is not as advanced as it could be though, as with this system, the player will only receive feedback after the time it takes for him to send the packet and be replied to, which can be too long to feel good for something like an action game. Therefore, one might implement predictive movement, where the client independently evaluates the player's input and immediately acts on it's own, but then, if the response from the server comes and it disagrees with the decision, the client revises its state to be equivalent to the server's. I haven't yet implemented such a system, but it's in my near future.

Also, here's a relatively brief article which explains basically the same thing, but includes information about the evolution of game networking philosophies, which may be of use to you:
gafferongames.com/networking-for-game-programmers/what-every-programmer-needs-to-know-about-game-networking/

This was a few years ago, but I've heard that Pygame was pretty outdated

I see, thanks for the detailed response. Some of this info I will definitely take into account.

And yeah my system is much more like the predictive movement system (when I get there). Each player has their own physics running and they transmit their position and vector, the server will check if it's in bounds and then send out that data to the clients. I'll need some interpolation to adjust positions and movements if they start to go askew.

You're probably right, I have never used it though. There is also pyglet as an alternative to pygame.

One thing I forgot to ask, do you send the entire game state in on lump or do you send individual states of each character to the clients?

idk i'm clearly retarded but if i set a limit rotation constraint on all the bones have it affect the y coordinate have the min and max set to 0, it does jack fucking shit.

...

found the issue the pole target was overriding all the other constraints and manual rotations.

Still remaking all UV maps
Kill me

This is why I don't manually model levels.

lesson learned

Hey uh… is U.V Mapping supposed to be harder than rigging? I used blender and seem to not have a fucking clue as to how it works.

UV mapping is easy TBH

what the fuck have i started.

What software are you using?

well, this is just a mapping tool, Violet UML Editor.
It's gonna be GM:S for the actual game.
Posted some in-engine things before, but I reached a point where i need to plot out the story and shit before doing anything else, hence this.

Does that UML Editor have an XML export feature? It seems to me that making an XML parser and then just taking in the output from that editor would probably be the easiest way to go about handling dialogue, if it's all going to be that complicated. If you try to notate all of it into some scripting language yourself I can guarantee you're going to make a ton of minor bugs and make things just overall 10x harder for yourself than it has to be.

Are you lolisimdev?

Loli sim dev made his own engine in C++, he wouldn't go to GM at this point

just use an XML lib
pugixml.org/
grinninglizard.com/tinyxml2/index.html
etc
there's loads of em

that is just TORposting, now this one is impressive

only .png or .html, it seems.
though i might have to look into that.

no. have not referred to myself as anything but nodev yet. besides, LDdev sounds weird.

I just have no fucking idea how to unwrap models without it take five million years and spitting out some ungodly deformed mess. Any advice?

there's different options on how to unwrap, i'd recommend just look for a youtube video on it.

i'm no pro but have done a bit in blender, basically don't use auto unwrap ever, unwrap from each side and make sure you get a clean outline of the side view when you make your uv sheet. it's really pretty easy once you've done it once or twice.

Waiting for Godot 3.0

well done

Where is the line drawn between a reference and plagiarism?

For example, if I wanted a cheat code in my game to turn a character into a Touhou, would it infringe on that, or just be a "cool nod"? Considering having my game be commercial, though.

Does someone use the pepe trading system atatched to bitcoin?

not terrible, not great, but I learned a lot… so that's a plus.

Animation itself is okay but that's pretty generic indieshit, user.

It is bad.

unless that's one character over another's shoulders, the knees are far too low

it's for practice, it's not supposed to look good (generic as possible palette, intention is generic as possible design) tbh.
focus is learning atm

Well then you should be learning a non-shit style, user. You won't learn dick by copying shitty art styles, anime proves this.

post examples

Is that an amputee in a burka, or is he running with his arms behind his back?

it has no arms, it's only legs and a head, kek

Literally anything pre-Indieshit era. Look into dithering, colour limitations, etc.

anime shit

hate those art styles, super ugly

ugly as it makes surfaces rough, unless you do super detailed sprites which take multiple hours to animate each frame, which I'm not going to do

tbh I already know the art style I'm going with, but I'm practicing until I reach that level as it requires a lot of knowledge of the "tricks" used to "imply" details.

Not surprising. Can you actually draw?

I know it has a cape and no arms, I just wanted to make a Naruto running joke

to answer your question
I've been drawing normal art since I was about 8, so about 20 years now, and was taught by someone with a bachelors in arts; who is actually a damn good artist.
I'm just sub-par at pixel art, bcs I'm still learning the "tricks"; which is the same with any form of art really.

sorta does look like when one of the akatsuki guys running ninja style with shitty coloring hah

go 3d and bake sprites with blender

and i have a 12 inch dick and if you ask for proof your a faggot

No, I just think you're a tad bit gay for insulting animations far better than yours as ugly. If you really want to learn about pixel art the truth is that most of the good stuff was drawn by hand first or 3D rendered and then redrawn in pixels. That or drawn on graph paper and scanned. Actually doing it dot by dot sucks ass.

the animations were great, but I mentioned that - to me - the 'artstyle is ugly (color palette is nice, but the style of drawing is unpleasant to my eyes).

The artist that I will be emulating definitely uses concept art, and I will too; as it's a lot easier to draw something on paper then just pixelize it.
However, I'm going for not too much detail same as the artist I like, as that artist uses a lot of tricks to "imply" detail so it looks amazing - to me at least - and it doesn't take hours to animate by hand due to said limitations implied via the style.

How serious/detailed do you think my grocery sim should be?

Because I used SS13 sprites, someone in another thread mentioned they had expected it to be somewhat similar in terms of tone and stuff, and that got me thinking. I did sort of intend for it to be like RCT in a way, but I dunno.

Sorry for the late response, I went to make dinner.

From what I've read, networking is too slow to be able to send entire game states unless you're on a local network. You can't transmit every relevant variable in your game to every client on every single update tic. In optimizing network performance, the name of the game becomes sending as little data as possible, or rather, only data which is absolutely necessary, and finding creative ways to minimize it. Here are some examples of good and bad ways to handle networking I've seen, in my extremely limited experience, from worst to best:

1. To send every variable constantly, and have the client make the game state from that. Similarly, I've seen server code which actually doesn't directly update a game state in the client, but just tells it where to draw every sprite on every single frame. This would be slow as fuck and probably break if done on the internet, because you're sending hundreds, possibly thousands, of variables per update tic. If I understand you, this is what you're suggesting by "send the entire game state."

2. Suppose you have an FPS with WASD movement. In this case, the keyboard and mouse input of the player is sent on every update tic, and then interpreted by the server, and the player is allowed to make their move once the server's response gets back to them. If they're holding W for 5 seconds, then W is sent on every tic until they stop. While player inputs are a much smaller amount of data than just every variable or sprite in the game, you're still sending data on every tic.

3. Suppose you have the same game as above, but rather than sending packets on every update tic that input is happening, only send packets as when something needs to change as a result of input. In this system, rather than sending W on every update tic, we instead send one packet telling when W is pushed, and one packet telling when W is released. Between these times, W can be assumed to be pressed. Suppose you have a 64 tic-per-second game, W was held for the same 5 seconds. You've reduced the number of packets he sent from 320 to 2.

Also, the number of packets sent increases rapidly relative to how many players are connected, because the reply is distributed to all players. If there were 4 players (including the acting player) connected to the server, then in system 2, then the button press would be distributed to 4 players 320 times, which is 1280 packets, for a total of 1600, all sent and received packets combined. This doesn't even take into account any button inputs that other players are sending. Only one player made that happen. In system 3, only 10 are sent in the same 5 seconds. The same basic idea applies to all data sent. Because you should be starting from identical game states, from that point, you should only need to send data which is absolutely mandatory because it directly updates the state of the game. No redundant or excessive data should be sent.

It is for this reason that I send input changes instead of the state changes, because there will always only be a handful of inputs, but a single input could make many, many changes to the game state, which would then all have to be sent.

In my game, player controls do not affect the player's character, and instead they send the unicode value of the buttons the player has pressed or released (as unsigned 8-bit integers) to the server. The server then interprets these as controls on its own side, and if it judges that they're valid (not walking through a wall or anything), it updates its game state based on them. Then, it distributes those same controls to all of the clients, including the person who originally made the input. In my game, all packets are prefaced with a numeric code which identifies what type of packet it is (input, in this case), and the ID of the client that it originated from, using the ID system I mentioned in my previous post.

After the clients received the packet and determine that it's a player input packet, the clients then use the ID of the acting client to look up, in the previously mentioned 2D client array, which character instance the controls should be applied to, and then applies them in the same way that one would apply controls for an offline game. Because all of the players are sent the same inputs for the same player, it results in the same game state, identical across all clients, assuming no packets were dropped.

If something goes wrong during this process, such as intense lag or dropped packets, there is the possibility that more data would need to be sent, in order to make sure that everyone is on the same page, which should probably be sent periodically, along with keep-alives, or something like that. Addressing this is something that I'll have to tackle soon.

imo it's best to plan only the foundation, and go from there (be pragmatic for what you should implement, imo); like do you want a multiplayer based game like SS13, with similar core mechanics?
Design your game around those core aspects.
Otherwise you end up with feature creep if you plan everything, and all kinds of extra "possibilities"; unless of course you have a large studio backing you up then go for it.
All that extra stuff can be added later, after you have that foundation laid, and you have fun "core gameplay" to work from.

Also, just make the game u want, and not what others want; there will always be people who share those interests/tastes.

I wasn't going to upload this until I was totally complete with the shoe part but I wanted to get some early opinions on my work so far. I'm also thinking of giving the Stands more realistic materials and textures but I realized that that doesn't make any sense since they aren't really made up of real world materials.

His eyes are too far part and he has no whites. You might wanna investigate ASB's shaders and stuff because they managed to recreate araki's style pretty good, and I know models have been ripped from ASB because I've seen mods use them. Also he needs to be more buff.

looks like my BIOS menu

I investigated a lot about the shaders in ASB, they made their own custom material and its pretty hard to even make something similar. The good thing is, most of the detail is actually in the textures,

if you are good at textures you can get something that looks pretty good.

I actually ripped some of the models let me see if I can find a texture

I found the DIO textures.
I had Killer Queen and The world too but I can't find them

Just get a dithering brush, doesn't take that long. Then have the two shades you're dithering be pretty close in color, not noticeable unless people zoom in. It takes more time for sure but it's only multiple hours for a single frame if you're doing each pixel carefully yourself.

I'll say.

Why does filesize matter?

Well 138 replies in I finally post progress on Kowloon in the Kowloon themed thread.

finally found this cunt
god damn

and the texture

Last for tonight.

Having infinite bandwidth isn't the same as infinite speed for that bandwidth. Even if you have free internet access until the end of time, webpages are still going to take a certain amount of time to load, and most people are stuck with a shit telecom who gives them shit speeds. The image in question, even with the mistake, is still only 28 KB, which in the context of modern internet speeds (and hard drive space), is in the realm of "who gives a shit," but it's an argument about best practices. It starts to really, really matter and affect load times once you're dealing with larger and more complex images. Your average internet user has the attention of goldfish and will leave sites very flippantly, and often due to load times.

However, for this image, there's really no excuse for it anyway, since even if you don't care about filesize, it blurred a perfectly good image arbitrarily. Of course, all this is assuming that the image was originally made by a direct screenshot from the game, and not person A doing a screenshot of a youtube gameplay video uploaded by person B.

Alright finally finished this one. Might touch up the hair/collar later but we'll see. Also figured out a new way to do my shading, and I'll just have to go recolor my older sprites.

I see we have a lot of "devs" here.
That means you should be able to solve this right?

You first.

...

kek looks like there are already some buttmad artists who can't into programming.

4chan/vg/ misses you

If you wanted help, you should have just asked for it.

I can show you my working solution right now but that would defeat the purpose.
If you don't enjoy programming then why are you in gamedev?

Who has time for such pointless problems?

To make games, not solve retarded /g/-tier programming baitjerks for /vg/ shitposters.

And how will you make games when you can't program? :^)

ah, the old "I'm not gonna spoon-feed you, you tell me" argument.
fine have some psudocode:
[code]
get string until end of line;
loop from 0 to string length{
take string[0] and compare to loop of from string [1] to string length;
if ch == ch cout and break
}
go to new line
[\code]

i swear it was \ not / last time i posted code

You don't learn to program. You either understand or you don't.

Oh my gosh user! Are you a hacker?! :0

That finds the first occurence of whatever you are looking at.
You are supposed to find the first non-repeating character.

Guess you are one of the people who don't understand according to the clever defintion of

A brilliant exercise. What's your next pointless endeavor? Capitalizing all the letters in a given file? Checking parenthesis in your code? Making a shitty spreadsheet calculator when you should just be making your game?

lol it's not plagiarism you have to worry about

it's copyright law

Why are you so defensive about your non-existent coding skills is the real question.

Fair use, just like Yandev can use Underememe models and other cancerous shit in his youtubebaitgame. Not to mention that almost all 2hu material online is made by fans because japanese law doesn't work the way it does here.
That said, using references like that makes you just as cancerous as Yandev.

It's 3am over here and I don't feel like refining my code any further.
now fuck off

#include #include using namespace std;int main(){ string word; char letter; bool match = false; cout > word; letter = word[0]; for (int i = 0; i < word.length(); i++) { letter = word[i]; for (int j = 0; j < word.length(); j++) { if (i != j && word[i] == word[j]) { match = true; break; } } if (!match) { break; } } cout

This looks fine, but didn't you mean
if (match){ break;}
?

I think the only thing I would do different would be using a goto rather than an extra boolean. Breaking out of nested loops is one of the few perfectly fine use cases for goto I can think of. Another one would be flow control in switches. In fact, in C# Microsoft forces you to use a goto for clarity, if you want to do a fallthrough.

Looks like it

cpp.sh/

Anyone familiar with a language's string library and basic loops/logic can do it. It would be more interesting if it had details to it, like you'd need to handle in a real use case. What if there are no unique characters? Are spaces a valid result? etc.

No.

if (match){break;}
will result in the program printing the first character that matches another character.

if(!match){break;}
results in the program printing the first character that doesn't match any other character.

The prompt asks for the second.

No, run the program and it's the opposite of what you're saying. Also use std::size_t instead of int for the for loops.

Nevermind, I'm a faggot who didn't read the question right.

It's OK, happens to the best of us.

cplusplus.com/reference/cstdio/
I believe these are optimal.

putc of course.

It does not read from a file provided from the command line as requested but i think we can let this pass.
Here is a reference implementation
open(DATA, "

Fuck you.

Fuck you, too. Don't supplicate to that useless piece of shit.

no bully

The wood texture looks a bit too shiny, you can't separate the materials so that the wood becomes less metallic?

Go for a zweihander.

looks sick! well done user

The in engine "metallic" is turned all the way down for that section of the material. I just have the roughness very low and no normal map for the wood grain so it looks very shiny. I was going for that "sanded smooth and covered in several layers of varnish" look like in pic related.

I will think about the zweihander.

how?

The World looks more baked than painted, with filters and masks doing most of the work, while Dio probably required a more clean layout for precise texture painting.

Fast blits.

What kind of use do the gun have? Is it a prop or is it used in combat? No gun meant for use is prepped that much.

Going to draw some concept art today. For mecha, guns, pilots, the main character, etc. A few questions:
1: What is AGDG's opinion on having cutscenes at the beginning and ending of every level; just short ones that 1: establish the mecha and give hints to its quirks and 2: give it some personality/a proper death?
2: How does the name "Augmented Light" sound? You play as a guy with some vision so augmented it'd make J.C blush.
3: I want to make the character very faceless (face never seen, masked) with only a basic backstory of "He/she was told if they made it to the highest point in the city and brought back the loot than their home district would be considered for assimilation into a greater empire". Is this a good idea?
4: How stupid does a gun that uses gas pressure to control recoil sound? The idea is that the Anti-Mech rifles fire such a hot load they're designed to vent the gas upwards and back to help mitigate the recoil.

1: Cutscenes are meh, unless your game is heavily focused on not being story focused. cutscenes takes control away from the player and ruins immersion, if you can do it while the player plays your game go for that instead. This is not the case when you won't to hype up a boss

2: good

3: Sure go ahead, but he shouldn't be 100% emotionless, give him a reason for being masked.

4: This is basically what modern rifles like the HK416 does, they vent the gas back causing the gas to mitigate the recoil and reload your weapon in a more effective way.

If the player skips it, he should still be able to play the game. Have all hints and shit be done via proper game design. If you want to use cutscenes to set up a situation and tell a story, that should be fine
The backstory is so small, it should just be given to the player directly. The whole game's premise seems to be around survival, so why need to give the player any more motivation than that? Just Like Kill Mechs

Checked.
The game focuses heavily on not being story focused. Every level wholly revolves around destroying a mecha of some sort. The cutscene would be thirty seconds and most and skippable. The alternative is shitty Half Life 2 style "interactive" (unskippable) cutscenes.
As for the mask the reason is that the air in the city can at times be incredibly toxic and it's connected to an aiming device of sorts. As for the rifle would it be realistic to show it venting a large amount of gas for every shot? It would be firing a currently impossibly hot cartridge.

Just a further explanation from my understanding of the system on the gun-
As the bullet passes through the barrel it passes by the gas piston, the pressure in the barrel, not the heat, causes the piston to push towards you and reloads a new bullet into the chamber. This + the gas that gets vented towards you during the reload compensates for some, not all ofc, of the recoil of the weapon, making it much easier to handle.

Yeah in that case, go for the cutscenes,
if toxic air gas masks are a given.

I think it would be within the ranges of what is realistic, it's not that visible here but it might works as a reference for you.

youtube.com/watch?v=6q7ysk0Nbmo

more about the cutscenes, show the mecha properly shutting off before playing the death one otherwise the player will think is just a phase change

Thanks for the advice. I should explain that I'm coming from the viewpoint of ps1 era games where you'd get a tiny skippable cutscene after a level. I plan on adding an option to disable movies entirely.

I definitely plan on that. Visceral and visible damage to the mechs is a must here.
youtube.com/watch?v=6q7ysk0Nbmo
Probably going to stylize it a bit for rule of cool. Thanks a lot though, I'd like to keep the guns somewhat /k/ friendly.

pixelation.org/index.php?topic=9173.0

Source engine maps are so comfy.

Thanks for the critique
I wasn't really trying to copy off the exact ASB likeness, I just wanted the structure of the model. The game is going to have it's own custom style I want something that's more realistic and less cartoonish. Unless of course people actually did like the All Star Battle style, which I personally didn't really find all too inviting.

And also I realized that putting the eyes too close together sort of made him look retarded, but it is something I need more work into.

Lots of powder I mean. Terminology like that is truthfully only further proof that /k/ is gay.
As for source maps I am going to say that there are few engines that handle vertical levels as well as source. Making horrifyingly complex and labyrinthine shit is actually not hard at all in that engine. It also has far better work flow than UE and Unity. If you just want to make a quick FPS there is nothing better. After I'm done making the exterior and "outside" of the city I am going to begin working on interiors. I want to achieve the E.Y.E teams mythical "so big it crashed the engine"

non gay board when.

Having a bit of trouble layering my shopper sprites while keepimg it flexible. Depth sorting isnt useful because the depth of say, the arm changes depending on whether the character is facing left or right.

Another thing is that I cant just draw the body then the shirt; if theyre facing right for example, I have to do left arm, left sleeve, body, shirt, right arm, right sleeve.

I can just draw that and it works but it offers zero flexibility and I have no idea how to fix that

How are you rendering?

I'm pretty sure /k/'s not gay so much as it's just willing to fuck anything.

I fucked around and got a triple double.

Thanks!

Im using XNAs SpriteBatch draw functions with deferred rendering. I tried sort ordering but ran into issues as I described… not sure how I can elaborate more

It's a fine idea and something that is actually used however it does have a downside
The amount of force the gun pushes back with is equivalent to the force the bullet gets.
That means if you're venting gas to push the gun it's not pushing the bullet or at least not as much so it should result in slower bullet speeds, now exactly how significant this if at all depends on the specific implementation.

The thing with the world is that his point on his mask is supposed to kinda morph into his nose.

eh not really most of the gas used for this shit is from the tip of the barrel after the bullet has gained most of its velocity more like an extra 3 in for the barrel or rather an attachment. that vents Excess gas backwards and up to produce a rocket like effect to mitigate the first of newtons laws. if it has any drawbacks it is that it adds weight to the gun weight to the tip making it off balance but if your playing around with a anti material rifu that is really more or less irreverent more taking up a upgrade slot more than anything else.

the bigger question is do you understand exactly how big an anti mat rifu is? you are not going to be able to run around with one. they have been and allays will be stationary while fired do to their scale. and one is not going to be able to use the french marshal art of running away with one. i think you need to rethink your theme. making more of a mech hunting game, witch is what i thought you where talking about when you first started posting. they are faster than you stronger than you and have more fire power than you, but you get the first shot (if you are lucky), make it count.

Christ user I really hope english is not your first language.

You can balance it with even more weight

Posting footage as of 3:21 PM Atlantic time.

Makes sense.

This IS a mech hunting game it's just a bit faster than you'd expect. If you've ever watched Armour Hunter Mellowlink or something similar you'd see what I am going for. The goal is to create an experience where you are constantly half-shitting yourself in terror running away whilst laying traps and trying to find nooks and crannies to squeeze into so you can get a shot off without being seen.

Main character is a ridiculous cyborg so this isn't improbable.

Also as a note I suck at making quality webm's. Here's a screenshot.

...

diminish the fog

I almost finished texturing the T-64 tank, which will be put into my threedee RTS game. Honestly I find modding a 3D game a bit easier then 2D one since then I don't have to do some weird ass offset calculation

I was thinking the air would be shot out closer to the base of the gun like near where the person is holding it
ofcourse that would be very dangerous since it shooting hot air right at your face
But if you're a mech that doesn't really matter, then again if you're a mech recoil doesn't really matter that much either

Love2D is alright for small games, but Lua's lack of a switch like statement without external tinkering in C would put somebody off.

It depends on what you want out of the language i suppose. Lua is a loosely typed language where everything is global by default, and python is more strictly typed. My impression of pygame was that it runs slower, if that means anything to you.

As for SDL/SFML, it doesn't really matter so much which you use, as long as you stick to it and don't flip flop. Again with performance stuff, I believe SFML gets the window open just marginally faster than SDL nowadays.

You don't know what a kowloon type city is, huh?

tHE FOG HSINES TOO MUCH. iF IT WAS GREY IT COULD LOOK OOD

hwut

As of now there are no real plans to diminish the fog. I may put the fog further in the distance later though when I have more than the first layer of the city done.
Also the fog is not shining it is simply coloured. I have set it to the colour of the lights as source has no support for lights colouring fog so I need to do this to add some realism.

It looks less like fog/smog and more like a very bright light bleeding into the scenery or obviously low render distance.

Jesus. That took longer than expected …
I now have contextual status hints for my HUD. One for general information, one for messages. one for warnings and one dedicated one for when you're on a ladder and pressing F would make you drop, so that you don't have to guess.

Fine, I cave. Better?

Yeah I think that's a better shade. Making the change to fog a bit more gradual would help too.

A lot better, but still not good IMO. Look how that building pops up out of nowhere at 0:15. Currently you have neither fog nor smog, but a wall. It should start closer to the player and then be waaaaay more gradual.

It'll be used in combat.
Changing The roughness won't be too hard.

Here, good?

Is there someone organizing a raid to clubpenguin?

If Lua doesn't seem to have some feature, that's just because you're not abusing tables enough. Switch statements can be easily emulated by a table with suitably chosen indices, e.g

Switch statement: switch(x)
case(1): functionA()
case(2): functionB()
case(3): functionC()

Lua: switch_table = {
1 = functionA,
2 = functionB,
3 = functionC
}

then call switch_table[x] when you want to execute your switch statement. Somewhat more retarded than a switch statement but works.

Should be switch_table[x](), sorry.

Nice, just needs a couple more textures/props to hide the repeating balcony textures.

Nice. Just realized you used the same music as this.

Looks better to me. If anything you may want to consider bringing the fog closer to the player so there's some slight haze even nearby if you want more choking/claustrophobic feeling.

I got lost trying to find the fucking brain in the suitcase in EYE, this game is going to ruin me

Jesus Christ I didn't either. How intriguing.

Trying this now.

I like mazes in FPS.

I find absolutely no useful information on how to do character controllers on the entire fucking internet. Im using bullet. The builtin character controller is buggy, almost entirely undocumented and everyone says not to use it. A simple kinematic character controller can't possibly be so fringe that there's so little information about it. Anyone ever written one?

i see ur using irrlicht eh?

Found a list of engines

Changed fog up again as different anons requested.

now narrow it down to only free software engines

Find one gameplay mechanic that's fun, and base all other additions on top of it to support that fun.
Keep asking yourself "Does this feel like the same game?" and "Does this support the core gameplay?"

Seems pretty weird that the one building looks so much higher than the rest, when it seems like the fog is there to give the illusion that the buildings are all much higher than you can see.

Maybe make the fog system depend much more on height than width or depth. That way the area you play in is visible all the time so you can have a strategy for how to tackle a room of enemies, while the buildings are just backdrop and thus can be hidden.

I will now hand you down the knowledge of my sensei, StephenLynx.

DON'T USE A CHARACTER CONTROLLER THEY'RE FUCKING GAY AS SHIT USE A RIGIDBODY

Eeeh this probably won't happen for two reasons.
1: No real "rooms of enemies" in this game
2: I actually cannot program things like this

Which? The one with the yellow lighting? I am currently trying to figure out how to mitigate this issue.

who's this?

if the yellow light has bloom that cuts through the fog, either eliminate the bloom or change it from being light to just a shadeless surface so the fog can obscure it. Meanwhile, add lights to the ground level. I feel bloom is too overused in modern games anyway.

What if bloom was used only on emissive surfaces rather than any light source?

Eg a computet screen

Make the player smaller

the guy who made lynxchan, who apparently lurks 8/agdg/

Ye, it's comfy so far

Rigidbodys have a shit ton of problems, they slide down slopes, they can't climb stairs properly and so on. Or do you mean some sort of complex rigidbody-on-top-of-raycasts thing that you see sometimes? Do you have experience with it?

i really haven't messed with stairs or slopes yet, so i didn't know this.

Empires looks like shit for a reason, user.

(Checked)
It's just an issue with the engine in general. Figured out a bit of a fix.

there's always ODE

but have a look at the capsule shape and friction settings

The idea behind using kinematic objects instead of rigid bodys for character controllers is that you have fine control over the movement, they don't slide or get bounced around.

Did you ever do anything with it? I like bullet physics so far, and my problem is more about the concept of a character controller, as in what sort of objects you use.

physics objects are meant to slide and bounce though

but you can use restitution to control bounce and friction to control sliding

any Holla Forums made pixel art that doesn't look like indieshit?

this is how my engine acts
rigidbody is a capsule with a radiuhs of 1 and a height of 2.

holy shit i'm retarded

And it works well? If you say it, i'll try it. Just don't want to spend a day optimizing a solution that could never work in the first place.

fuck if i know bruh

That looks good, no sliding at all, precise controls. And that's just a rigid body? How do you move it, applyCentralForce?

forces will act like acceleration
setLinearVelocity will behave more like a typical FPS

there is sliding, stairs work, i use applyCentralImpulse.

webm related is applyCentralImpulse, somehow much faster than applyCentralForce, i had to reduce the force to a tenth.
I've got a few more questions: How do you keep the capsule from falling over? What sort of friction settings do you use so it stops immediately and does not accelerate but instead keeps a steady pace?

My motion state is:
btDefaultMotionState* groundMotionState = new btDefaultMotionState(btTransform(btQuaternion(0, 0, 0, 1), btVector3(0, 0, 0)));
The mass is 2.
My gravity is (0,-29,0).

Thank you very much user. How much force do you apply and how do you keep the thing from falling over?

Removed a few shirt symbols, added speech bubbles, now they can think of things I guess

setAngularFactor

maybe set to default motion state every frame?

wrong

...

Thank you. Even with your gravity and mass settings though, the capsule is floaty as fuck and doesn't stop once the applyCentralImpulse stops. Also the thing gets stuck if a wait a few seconds after spawning it above ground, and i can't move it anymore, i had that problem from the start, any ideas?

Maybe apply some sort of friction? I don't know shit about what you're doing, though

6/10 objects remapped

set mass
set friction
set restitution
etc

when you let go of movement keys set the velocity back to 0 if you want the object to immediately stop moving

bullet is a huge library, you just have to play around with it

Added socks, sandals, shoes, boots, heels

>no socks and sandals
>no toeless socks
Good job though user, the heels look kind of weird from a side angle though, like they're too short or something, but the rest of the shoes seem good.

Actually, I'm working on the underwear right now. Socks and shoes are separate, so this will be possible.

Is anyone here familiar enough with Unreal Engine to who knows how reverse the direction that Bump Offset shifts the UV's?

I'm working on webm related but I need the circular mask to shift in the opposite direction to how it's currently moving.

The bottom ones are leggings, so you can combine them for long underwear, or as chaps, or as pantyhose.

Hey fags, gonna do it 4free
The devs making Angels Falls First, an FPS Scifi Battlefieldesque game that has a small following here, are hiring for U3 mapping and 3D artists.
Also USscrip

If any of you autists are interested in getting paid over working part time on the weekends and gaining dome pocket change, heres the link:
steamcommunity.com/games/367270/announcements/detail/630905083630402881

why are you speaking in third person

Fuck, they found out
seriousy, I just like the game, its the closest fo Battlefront 3 that I'll get

then go work on it

most people are here because they want to make their own games, not work on someone elses

We're gonna make it fam

Don't have the requirements nor am I interested on doing so
Fair enough

I've seen that there's a paradoxical process which happens at an individual's initial engagement with a skill.
As far as I'm aware, the only way out is to care so much that you force yourself to actually find the time/motivation/self-discipline needed to sit down and grind like fuck on things that you don't care about as much as whatever magnum opus you think you have kicking around in your head, until you can make what you really want. If you cannot do this, you will not make it.

The second, and maybe better way is to stop being a perfectionist and throw yourself at the game as soon as you reach Step 4. You can't make a game if you're afraid of your game, just try to make it and see what happens.

I'm putting on the back burner for now. The way I want to do it may not even be possible. Instead I'll jump back into modeling. I decided on a mace.

looking good
now add a yellow pissfilter and you can claim to be AAA dystopian cyberpunk producer

(1/2)
Don't listen to .
I've been using both for an extended period of time. While I was still using Unity, I made a rigidbody based player controller. Now that I switched to Unreal I'm using their ordinary character controller as a base. Both have their problems. Making a good character controller can become rather difficult since you have to account for a lot of bullshit that the players may or may not do.

A rigidbody based controller seems like a good idea at first, because things like interactions between the player and other physics objects, jumping and falling are already taken care of (at least in large parts) for you by the physics engine. In fact, the only thing that's a bit more involved from these three things is the first. You absolutely are not allowed to touch the velocity of the rigidbody directly. If you do, you break physics interactions between your character and the world. Imagine you set the velocity to (0, 0, 0) when the player doesn't press any button. You literally turn the player into an immovable object. Even a 500t object pushing against him will come to a complete halt immediately. The opposite occurs, if you move the player by manipulating the velocity vector directly. A 500t object that is supposed to stop the player, will just be pushed away like it's mass is 0 since you overwrite the players velocity immediately after he has been stopped. The player will have become an unstoppable force.
mentioned slopes and stairs as problems. Those are cases you'll have to take care of regardless, since neither controller plays nice with them by default. Either way you have to define a max. obstacle height, then detect them as the player collides with them and check, if you can move him on top of them automatically. Slopes are relatively easy. The way I handled them was by massively increasing the friction and/or inertia of the player when there is no input.
A big issue (and the reason why I won't touch rigidbody based character controllers ever again) is that you tie your character to the physics engine. Depending on how your physics engine is implemented, it may run at a different framerate than the rest of your game. In Unity that meant that the player was (by default) capped at 50fps. It caused visible lag when walking sideways and rotating the camera, since the position would only be updated at 50fps. In the end I had to detach the camera from the player and write a dedicated script to make it follow him on a smoothed path. This introduced lag (the camera was always behind a bit), but it was small enough to not matter that much (not making a precision platformer here).

(2/2)
"Ordinary" character controllers have different issues. The first ones that come to mind are that you have to take care of jumping, falling and your player applying force to other physics objects yourself. Jumping and falling are simple enough, but what's a real pain in the ass is turning your player into a pseudo-rigidbody that has most of the benefits without any of the downsides.
Since I use Epic's character controller as a base, I haven't touched any of these and can't tell you how exactly they did it. However, what I can tell you is that Character.h and Character.cpp are 894 and 1573 lines long, respectively. You can find the relevant API documentation here:
docs.unrealengine.com/latest/INT/API/Runtime/Engine/GameFramework/APawn/index.html
docs.unrealengine.com/latest/INT/API/Runtime/Engine/GameFramework/ACharacter/index.html
docs.unrealengine.com/latest/INT/API/Runtime/Engine/GameFramework/UPawnMovementComponent/index.html
docs.unrealengine.com/latest/INT/API/Runtime/Engine/GameFramework/UCharacterMovementComponent/index.html
This might give you some idea of what to do. (If you have a GitHub account you could also check the source yourself.)

The benefits of using such a character controller is that it runs with the main tick of your game and thus scales with increasing framerates. It also gives you very fine grained control. Here are some pieces of functionality I deduced from what I did with my base class. You need to take care of:

Hope this helps you at least a bit.

Working in UE4, best way to detect a change in a variable? I already set it up so it has to be modified via setter function, but I'd prefer not to have to go back and put in a call to every single thing that needs to be triggered when health changes so it should be something others read from it rather than the setter calling stuff. Flip a boolean for one frame? Custom event that is called (which I then can tie up in Blueprint too)?

I could probably use RepNotify too, but that depends on if UE4 will throw a shitfit because my game is not multiplayer/networked and causes problems. Apparently it does not run on servers, but no clue if that's just dedicated servers or listen servers too (latter is what I assume it would be treated as in a singleplayer game).

Looks a lot better to me, you can play around with the gradient/fade/whatever a bit, but seems much better than the rather sharp white-ish wall of before.

What you want is called a dynamic multicast delegate. (docs.unrealengine.com/latest/INT/Programming/UnrealArchitecture/Delegates/) Think of it as a list of functions that you can call. I'll post a specific example in a minute.

Here is my delegate that allows outside classes to react to when the player's health changed.
// PlayerCharacter.h#include ....// This only declares the type (dynamic multicast), name (FPlayerHealthDelegate) and signature (float, float) of the delegate.// It's NOT the delegate itself.DECLARE_DYNAMIC_MULTICAST_DELEGATE_TwoParams(FPlayerHealthDelegate, float, CurrentHealth, float, Difference);UCLASS()class ....{public: // This is the delegate itself. Think of it as a list of functions that other classes can add themselves to from the outside. // It's a subscription. Rather than let everyone check whether the health changed themselves, you let the character (i.e. the one who knows // when it happens) know that you want to be notified when it does. /** Fired whenever the player's health changed. */ UPROPERTY(BlueprintAssignable, Category = "Events|Character|Health") FPlayerHealthDelegate OnHealthChanged;protected: // This merely exists for reasons of encapsulation. /** Fires the OnHealthChanged event. */ UFUNCTION(BlueprintCallable, Category = "Character|Health") void HealthChanged(float CurrentHealth, float Difference);}

// PlayerCharacter.cppvoid APlayerCharacter::HealthChanged(float CurrentHealth, float Difference){ // This iterates over the list and calls every function with these two parameters. OnHealthChanged.Broadcast(CurrentHealth, Difference);}

Subscribing from Blueprints is easy.

Subscribing through C++ can be done like this:
// Whatever.cpp// BeginPlay() is the EARLIEST you're allowed to do this. Don't try to subscribe in the constructor!void AWhatever::BeginPlay(){ APlayerCharacter *Player = ...; if(IsValid(Player)) { // Tells the object (Player) that its delegate (OnHealthChanged) should call the function (AWhatever::IWant....) of the object (this) // whenever its health changed. Player->OnHealthChanged.AddDynamic(this, &AWhatever::IWantToKnowThatTheHealthChanged); }}// This signature has to be the same as defined by the DECLARE_... macro in the header!void AWhatever::IWantToKnowThatTheHealthChanged(float CurrentHealth, float Difference){ // Do whatever}

it's a hobby bro, not a religion, you don't need "mastery"

most pro developers don't even have mastery

Should have put GNU+Linux on that laptop.

Then made game in Love2D with lua.

ask me how I know you don't know how to use bullet.

As I said in my post: Unity & Unreal which use PhysX.

unity also supports bullet

So if you're working in Unreal, why are you lecturing someone on the use of the character controller class in Bullet? They're nothing alike as far as I'm aware.

Yeah, Unity supports a lot of middleware. I wasn't using it though.

Gotta admit that I didn't read all of Anons post, so I didn't know he was using Bullet.
However, aside from the possibility of it being useful for other Anons too (many of us will have to deal with player controllers at some point), Unity and Unreal have those problems for good reasons. It's not like Nvidia just chose to make PhysX run at its own rate — with all the problems that this brings — just because they kinda felt like it.
If Bullet works differently, then most of what I said obviously doesn't apply, BUT you're bound to have other problems that come due to the trade-offs that the Bullet devs chose. Realtime physics simulation is a complex topic and there is no "perfect" solution as far as I'm aware.

how do I pass a function as an argument to be used in a function in c++
Like with std::sort you can pass a function that will be used instead of the < operator

either use a function pointer or std::function

you like do realize you can simulation speed in bullet based on the delta you pass to stepSimulation in bullet my dude

W-what?

you can pass a scalar timeStep to stepSimulation with a bullet dynamics world, and this allows you to adjust simulation speed on-the-fly.

AFAIK

You are correct, the bullet website tells you how to adjust to even 120fps.

what's the syntax here?

#include #include int main(){ Google g(); std::cout

function pointer is &function_name
std::function is std::function
to instantiate and to create is [how you want variables to be used in the function from the outside scope](arguments declarations){ function code}

Thanks user, suspected delegates might be the answer and you put it a lot clearer and concisely than the wiki does.

I already know how to sort

thank you
for anyone interested it ended up looking like this.

f2 closestUnit(f2 position, float radius, vector teams = vector(0), int currentTeam = - 1, f2 exclude = f2(- 1, - 1), string name = "any", bool visible = false, relation rel = ANY, const std::function& f = [](Unit* a) {return true;});

actually that lambda isn't quite right it should have been.
[](Unit* a, Units* b) {return true;}

i have a c++ book right in front of me:
#include using namespace std;void function_accepting _pointers(int some_data, double (*pf)(int)); // the function thats the pointer must return a double and take in a intint main(){int x =20;function_accepting_pointers(x, niggers);return 0;}double niggers ( int N){return N * 14.88;}void function_accepting _pointers(int some_data, double (*pf)(int)){cout

hmm does anybody got a suggestion for a tier 0.5-1 light shell/mortar artillery vehicle?

...

120mm Mortar on a M-133 'Gavin' huh? seems to be a little big for Tier 0.5/1 tbh
ayy puta

you could chuck a 65mm mortar in the back of a humvee or something

i was just suggesting something light armor with a mortar or rocket launcher in the back

Not this meme again.
Assigning velocity is no different from calculating the exact force required for the object to reach that velocity. F = mV At least that's how PhysX appears to handle it.
The problem is that video game movement is not realistic, and nobody actually wants it to be realistic.
It takes a lot of force to move a character in a precise fashion, but it isn't an infinite force.
The issue of the player being an immovable object can be solved by handling collision events yourself. You might as well kill the player or ragdoll them for a moment if they're hit by something really heavy.
Either way, don't rely on physics for critical gameplay elements because not a single physics library is reliable enough to do that.

here is your (you)

does she move the ears on purpose or is it a glitch?

Hmm yes that sounds good, I will make it soon then.

It's not a glitch!

Now that's interesting. Which version of Unity are you using? I'm asking because Unity has updated to PhysX 3 (which was was a complete rewrite IIRC) in 5.2 or 5.3.
The experiences I made were all the way back during the closed 5.0 beta. And I swear to god, that's how Unity behaved back then. I know, because back in the day I had to deal with this crap after the Unity dev in charge of creating the sample assets gave up on the rigidbody based character controller. He couldn't make it in time since there were lots of issues (in particular the interaction between the player and other objects).
I took it upon myself to fix this because I needed one. After bashing my head against a wall and trying various approaches, I finally had a breakthrough and send my solution over to him. I still have the conversation in my inbox in the Unity forums.

Am I retarded or are memory errors literally unsolvable
GDB is giving me a stack trace that can't possibly be a problem but it's calling malloc
And apparently with memory errors stack traces are pretty much useless cause the actual problem can be in a totally different part of the code FUCK ME

you are probably retarded but gdb can be unintuitive. create breakpoints and use gdb to step through your code until it segfaults.

I know how gdb works man but it's not failing on the line that is actually the problem
Also it's doing several thousand loops before it actually fails so going through it step by step isn't really a solution

gdb is never late, nor is he early, he breaks precisely when he means to

Have you ruled out compiler optimizations? Debugging when they're turned on can be quite the fun experience, with the compiler modifying your code during compilation.

gdb can go fuck itself

I don't compile with -o3 and -g at the same time.
also valgrind gives me literally tens of thousands of lines of memory errors.

maybe you should learn to use the software and stop blaming the tools for your mistakes.

im not blaming tools for my mistakes im blaming them for not helping me fix my mistakes which literally what gdb is for.

That's pretty cute.

5.1.3p3 as always, no upgrades for me
My fixed update is 0.0083(3), which is about 120 FPS.
Another way to do this is by using the Rigidbody.MovePosition method, which also handles physics correctly while letting you move in the variable updates, but it introduces visible jitter when sliding against walls. You can counter this with doing a sweep&project on your own in the same update, but that'll make the actor stop when hitting just about anything.
I use the aforementioned variable update & sweep test approach in the actual game to keep things as smooth as possible, but that means no solid rigidbodies. (not that I need them)
But if I did need them, I'd have two "solid" layers instead of one; one for static objects and one for rigidbodies. I'd sweep against static objects, but not against movable ones. (or sweep only against dynamic objects that are 4t or heavier)
Or, building upon the above idea, you could keep them on one layer and project the velocity while using the mass difference between the actor and the object as a weight. This is the slowest solution, however.

The solution I came up with (first person controller, mind) was by using AddForce() with ForceMode.Impulse. Based on my current velocity and input, I calculated a force vector that accelerated me in an appropriate way. ForceMode.Impulse ensured that the physics were all sane, but it had the downside of not letting you turn rapidly. It was like you if were on ice. I fixed that by separating the turning and "walking" logic, if you wish. I split up the input and velocity vectors into their respective components and handled the forward and sideways axes separately. This force vector was applied to the player and the physics engine took care of the rest. Fast turning was done by keeping track of how much you moved your mouse and using Quaternion.AngleAxis() to rotate the velocity of the rigidbody around the up vector of the player. This didn't break physics because it didn't change the length of the velocity vector.

Well, I'm just glad that I don't have to deal with most of that bullshit anymore. Epic included pretty good character controllers in Unreal 4, so most of the annoying work was already taken care of for you.

Actually, scratch that post. That gif isn't even MovePosition, I've been using Translate the whole time.
Which is even more odd because Translate should break things, but it still works. I even changed the collision type to discrete just now, lowered the physics framerate, and it still refuses to break.
Also, setting velocity to zero will make an object quite stable, but not immovable. The mass still matters.

Which is completely redundant, given the F=mV equation. Unless you cap the acceleration, in which case you're introducing lag and floatiness to your controls. This is more realistic, but it often feels like crap.
If you're gonna do all those workarounds and play whack a mole with all the edge cases, you might as well just handle as much as you can manually.

And I've just implemented the weighted sweep, works as intended.
The projected velocity vector is lerped with the original using the ratio between the object's mass and the actor's mass.

Make this apply to enemies as well so you can fling them left and right like ragdolls.

you probably have heap/stack corruption

You know what just occurred to me?
You're just a random polish user and your animations are smoother than the ones in Mass Effect Andromeda.
What was their budget again? Something like $50 million.

Also what said. Few things are as satisfying as letting enemies fly through the air as you kick their asses.

also anybody got some dnb

40 mil, five years, aaa studio.
this is what happens when you prioritize diversity.

Try to create the smallest piece of code that has the memory error.

Create a suppression file assuming these errors aren't your own fault.

...

DnB?

yes but how do figure out where it's coming from

The issue is I have literally no idea where it's coming from also it's seems pretty random so I can't even reliably replicate it.

Making some interiors. Is the music chinky enough?

Oh fug, user. That looks really good. Are you making a cuhrayzy? The girl is adorable.

yes

Did you make all those assets? I'm quite impressed. What's the setting?

use compiler options or dick around until it works

i'm not happy until you use this

are the wings and stuff like some sort of devil trigger? or why does she appear differently here

What about a tile-based turn-based CTF game?

Christ, the 1 bit clicker jam meme is up to almost 350 participants

I have an itch to make maps so what is a nice fps to mod in linux?

red eclipse, xonotic, unvanquished, warsow, assault cube

and which one of them would you consider the easiest to do with blender?

i believe half of them are on the quake engine and use BSP maps and the other half use the cube engine using it's own map system.

im unfamiliar with making maps in blender. However, what kind of maps do you make?

normally I don't do maps only renders, but i want to try it because of this

...

No, you pick them up at one point, like almost everything.

The setting is anime a high tech, low scale setting where many people live in remote, self-sufficient cities.
Once of such places has fallen to a disaster a long time ago, and it's thought to be abandoned; you're sent to investigate strange activity there. and then skeletons pop out

why would anyone that has tasted the wonders of linux do that?
I'm just saying that i feel the itch because of that post

sorry, i misinterpreted

weirdly i've started using windows again to work on a game, and even my casual usage has increased
this is clearly not ok

One option, if you won't be using more than a few distinct functions, is to use templates. You can put your function in a struct, declare it as static, then put that struct in the template parameters (the type itself, not any particular instance) and then invoke the function statically. This has the advantage that the compiler is pretty much guaranteed to inline the function, while if you're working with function pointers, the compiler can't do that, resulting in slower code.
Example:
[code]
struct WrapperStruct
{
static void doSomething() {do things}
};

template
void myFunction()
{
T::doSomething();
};

I solved that by building a security/media server with spare parts, that way i always have a linux machine nearby, even if I'm gaming on windows

Every time I've had a memory error, the solution has been to read my code until I realise where I fucked up. GDB is not very useful unless you can figure out a particular address that does not hold the value you expect, put a watcher on it and run until something pokes it that is not supposed to. Then you get to figure out how the code that did it ended up doing it by examining it's relevant state and figuring out where it went wrong and so on. It's the worst kind of bug there is, but it can be solved.

...

Deus Ex remastered is looking good!

I quite like that; to the point and self contained. The setting for the Kowloon game so far is a anime post-cyberpunk world wherein humanity became so technologically advanced they said "fuck that shit" and accidentally wiped everyone's memories after murder-fucking most of the populace. One of the devices that murder-fucked humanity is rumoured to be at the highest point in the world-spanning city. Acquire the MacGuffin and get paid you are told. And then a mecha/cosmic horror popped out.

i really just want to run vs under wine, but that's never going to happen

And one more thing. At one point I had some sort of bug I couldn't figure out, don't even remember what it was. I found one recomendation was to use -Wall -Wextra -pendantic -ansi. With cmake this is done with set(CMAKE_CXX_FLAGS, "${CMAKE_CXX_FLAGS} -Wall -Wextra -pedantic -ansi") in your CMakeLists.txt. This will make the compiler as bitchy as possible, meaning there's a good chance one of the warnings will actually be the problem. You can alsy try using llvm/clang. I did that once and it spit out a huge number of errors and warnings, most being forgetting to insert default cases in switch statements. After fixing everything clang complained about, the bug was gone. Again, I don't remember what the problem was, I only remember fixing it this way. If you're using cmake, switching compiler is trivial, just do
mkdir Debug-clangcd Debug-clangCC=/usr/bin/clang CXX=/usr/bin/clang++ cmake -DCMAKE_BUILD_TYPE=Debug ..make
Then for release build, same but with Release build type (and in a different directory. You should always do out of tree builds).

dialogue system, npcs, and 2deep4u story when?

>making your porn browsing freedom machine as vulnerable if not more than a standard windows machine
real talk: Is just for dx11 stuff, and even that is just dmc4se

just pci passthrough it bro

...

was about to say that i don't have the right adapters to use a second gpu(i run an hp z620), then i realized my game was using opengl 1.1 and i have a ton of old gpus

how many pepes does this man needs?

you might want a second monitor too and don't be shy about putting the beziers in the middle

...

meme

When I get a programmer.

but I have thousands of lines of code and no idea when I introduced this memory error.

this is what the compiler spits out with those commands
pastebin.com/bn12ZNQd
there's actually more it just doesn't fit in the console.

So get to work. A quick scroll revealed that a lot of errors are due to undefined stuff and lacking -std=c++11. The former is because you don't include the stuff you use and just rely on it getting cascaded through other includes, especially std stuff. Include everything you use in the files you use it in. The second is a compiler switch you need, just like the -Wall I mentioned. Do these things and you'll cut out a lot of stuff.

rip this webm size sorry, but working on some tribes style controls

Looks kind of ugly, consider changing your filtering

how can they pull stuff like this with basic software. I don't feel so confident anymore

time and effort

what engine is that? something cube/sauerbraten based?

yes, red eclipse

damn i've fucked around with the red eclipse editor and never made anything near that.

the effort to make those arches and curved surfaces in that engine, insane

Dude.

>F = mV
Technically, no. Mass times Velocity equals Momentum. Force is the time rate-of-change (or derivative) of Momentum, and since Mass usually remains constant, that gives you Mass times the derivative of Velocity (which is Acceleration). Ergo, F = mA.

Of course, like you said, games don't actually simulate physics realistically, but do just good enough of a job so it feels somewhat natural.

You could make a Rocket Buggy. They were the GLA's light artillery from CnC Generals.

en.wikipedia.org/wiki/Moiré_pattern

Not quite the same

Last for tonight.

Gosh those club lights are seizure-inducing, but the fog is definitely better now. What do you have in mind for the HUD?

Hey, thanks. I'm learned a lot about the way source handles fog and some neat rendering tricks whilst I changed that up. Was the comment about the club lights advice?
I'm thinking minimal hud to be honest. Green text computer terminal inspired look with everything displayed alphanumerically; nothing fancy aside from the green look and blocky font.

* I've

I'm really digging the aesthetic this is taking on.

It's an extreme nitpick, but no club would be designed that way, with an area for dancing immediately adjacent to what's essentially a pit for its patrons to fall into.

It's Kowloon inspired, I don't gotta explain shit. Unless that's an honest criticism and not just shitposting.

What do you guys use for texturing work on Linux? Gimp? Krita?

The perfect remedy for overpopulation, constant dangerous situations in everyday life.

almost time for a new thread~

No, it was legit hurting my eyes. Source lights have a number of preset and custom blink settings. Perhaps experiment with the settings, and find one that's more in-tune with the club theme (and less of an alternating flicker seizure maker). Is it supposed to sync with music, or be a broken fixture (perhaps both)?

Hammer has some interesting entities to experiment with, and I have fond memories of making levels with it in the past; though I haven't touched it in quite some time. Something I recall: DON'T CARVE SOLIDS WITH HOLLOW BRUSHES, it crashes. How has Hammer been lately? I recall it being a pain in the ass to get running ever since they switched over their file system to whatever the fuck it is now.

Something you might consider experimenting with, since you'll be running from a mech and all and likely relying on audio cues, is env_microphone (it has some interesting uses). Also, are you using any soundscapes at all?

New one ready to be published at any moment right here.

still don't know if the dude just increased the amount of frames in a cycle or something else

Was it? I might have a copy of it on my flashcart, will check tomorrow.

NEW THREAD

depends on what you are doing, if it is realism import photographies in gimp then correct them, make seamless and generate bumpmaps

for cartooney stuff i paint some clues with blender then finish on krita

I plan to learn inkscape for logos, but since i haven't used it much i can't recommend it