next »

[tjcbs]

I'm working on a glowing trail that follows the mouse cursor.

SWEET BABY JESUS this is hard to get right. I thought this would take me like an hour or so. NOOOOO SIR. Its been days and days of trying and scrapping one approach after the next, mounting frustration, despair looming larger and larger. If I knew it would be this hard, I would never have bothered. But now I have to see it through. I *think* I finally see the light at the end of the tunnel. But good god, this sucked.

I give up, its just not worth it. What a WASTE OF TIME.

[SolarLune]

Sorry to hear that. Although, sometimes you're right, and you have to just abandon certain approaches if they prove to be too time-consuming.

Have you tried making a plane that's subdivided several times, moving the leading vertices to the mouse cursor, and making the following vertices trail behind that position by a few ticks, like a standard weapon or movement trail?

Have you tried simply spawning particles, or additive blended planes where the cursor is, and having them fade out or get smaller over time? At a high-framerate and without moving the mouse cursor a lot, it's an effective approach, I think.

[tjcbs]

Have you tried making a plane that's subdivided several times, moving the leading vertices to the mouse cursor, and making the following vertices trail behind that position by a few ticks, like a standard weapon or movement trail?

Have you tried simply spawning particles, or additive blended planes where the cursor is, and having them fade out or get smaller over time? At a high-framerate and without moving the mouse cursor a lot, it's an effective approach, I think.

The latter (planes) is what I was doing initially, and it looks *ok*. I don't think the former really solves the basic problem: the generation of glitchy, crap-looking junk under arbitrary, jerky mouse movement. And really even when it was working correctly it didn't look especially spectacular. There are much more important things I need to worry about right now. I may come back to it at some point though.   

[netravelr]

Right now I'm programming a number of things, but I'm currently working on an in-game level editor in Unity for platformer games.

[Klaim]

Recently I finished what I call a "hackable" component system (in C++) in a concurrent context.
Basically, it's a access-only-by-id kind of cluster of component pools which are not stable (they can be reshuffled in memory if necessary, this is for specific reasons, I have other kind of component system but they can't do the same "tricks"), so basically you push lambdas into the cluster associated with ids (or a predicate) to do something with the components. Of course it's for a context where there is almost no hard dependencies between components.
I use type-erasure so there is no other requirement on component types, they don't need to inherit from anything, updating is done by specifying an object that does the update all objects of the same component type for each cycle (or more).

The "hackable" (my game is more or less about "hacking", whatever the meaning) part is that I can declare that a child type of a component type, which implement the component interface if it is virtual, can receive operations/commands done on the parent type. This is a very game-specific feature which allow me to have components that "shadows" other components and react to the operations being done on them.
It's actually very strange and surprising to implement.

Now I need to write some tests to make sure that works correctly and then I'll have to make the client code of my game use it (instead of the current system which works but don't allow the shadowing thing).

[Blockers]

i've been working on some kind of fighting game prototype this past day or two.

its coming along better than expected

[oahda]

Nothing. I've been at work for more than two hours, and I've done nothing. God help me.

[Layl]

Terrain generation!

[tjcbs]

Terrain generation!

how fun!

[MW0]

I'm working on camera smoothing for a 2D platformer. The smoothing is done but now the player jitters due to rendering  .

[Boreal]

I'm trying to design a bytecode for a Scheme implementation.  At first I thought it would be possible to implement it using reverse Polish notation (just flipping around the expressions), which would allow me to use a stack VM (much simpler), but the fact that procedures can have side effects killed off that idea.  Otherwise the side effects would happen in reverse!

Man, I really love Lisp but I hate that it isn't a pure functional language.  Haskell is better in terms of basic language design (monads are heavily supported in its syntax) but I hate how obtuse the syntax is.  Maybe I'll just develop my own pure Lisp dialect and use that.

[Rainforce15]

Man, I really love Lisp but I hate that it isn't a pure functional language.  Haskell is better in terms of basic language design (monads are heavily supported in its syntax) but I hate how obtuse the syntax is.

same here, and one of the reasons why I always take a step back when I hear functional programming.

[aleios]

Working on my deferred renderer. Got it up and running along with the HDR pipeline. Now to profile and optimize Thankfully gDEBugger makes that process a lot less painful

Not many plans for it in terms of actual game development but it's a nice little side project which allows me to learn GLSL and push for better performance, less state changes, and better ideas from OpenGL.

[coppolaemilio]

I wish I was programming something, I'm refactoring.... 

[ThemsAllTook]

I'm equal parts grumpy and happy about this, so I guess it goes here...

My game development framework has been in need of an update for a while, particularly the iOS bits of it - last I really touched them was in the iOS 4.3 SDK, and it turned out my stuff wouldn't build with a fresh Xcode install with a 7.x SDK. After updating everything, I found that my unit tests would crash with an illegal instruction error when linked against the new SDK. I'm doing some things that are unorthodox in Apple's warped developer world, in that I don't actually use Xcode for anything - everything is built from a makefile, and iOS unit tests were run on the command line with some dyld environment variables set to make it act like it was running in the simulator.

The illegal instruction was a complete brick wall, since I was crashing before getting to the first line of main(). Very little I could do to troubleshoot. After trying everything else I could think of and stumping all of my Apple developer friends, I had the crazy idea of copying the old iPhone 4.3 SDK to my new computer and telling dyld to link against it when running unit tests. Lo and behold, it totally worked!

So, now I have to keep this dusty old SDK around just so that my unit tests will be able to run without crashing. If I ever need to link to some framework that's new or changed in iOS 5+, this will no longer do the job, so it feels very fragile. That's why I'm grumpy about it, but I'm happy to be unstuck and productive again.

[MorleyDev]

same here, and one of the reasons why I always take a step back when I hear functional programming.

Honestly, my favourite 'functional' language is Scala. I find it has the most readable syntax of all of them, and the most actually-useful (well, toss-up between Scala and Erlang I guess), but it does the least to "enforce" purity since it...well, allows for variables. And allows for using a lot of Java libraries, which are obviously not particularly functional. Which is fine for me, I started with C++. Programming by discipline is kinda engrained in my bones by now.

Currently working on my C++ Dependency Management tool. Wrote the basics for my university project, now time to make it actually useful for more than the simplest use-case. Then comes making a web front-end for the ReST API (this API contains the information about the dependencies the tool uses), and a small public release. Client is written in Scala, the ReST API in Javascript/Typescript (node.js).

[Raptor85]

got bored working on 3d stuff, took a few hours and implemented simplex noise (with modifications and a second pass function to make the resulting terrain more suitable for simple isometric layers) and a simple isometric renderer


then i got bored and added hardware accelerated pixel plotting and streaming textures using pixel buffers to my engine, for no particular reason other than it sounded interesting and i was wondering how much faster it was than using something like SDL.

[poohshoes]

We are using OpenTK to try to support as many joysticks as possible.  Joysticks are only discovered when OpenTK is initialized so if you plug in a joystick after the game has started or if a joystick get's disconnected during game play then you have to restart the game.

Yesterday I hacked together a solution by just exposing the discovery code and then in the games code I call it every 2 seconds.  This is fine and dandy but I appreciate access to the source and I want to be a good open source citizen so I am trying to figure out how to fix it properly so nobody else has to deal with this in the future.

[ThemsAllTook]

We are using OpenTK to try to support as many joysticks as possible.  Joysticks are only discovered when OpenTK is initialized so if you plug in a joystick after the game has started or if a joystick get's disconnected during game play then you have to restart the game.

Yesterday I hacked together a solution by just exposing the discovery code and then in the games code I call it every 2 seconds.  This is fine and dandy but I appreciate access to the source and I want to be a good open source citizen so I am trying to figure out how to fix it properly so nobody else has to deal with this in the future.

Just in case you need another example to reference, I've written an open-source gamepad library here. Feel free to copy any portions of it that might help. It uses the same strategy for discovery as you described, in that it detects once at initialization, and the client code can call the detect function (Gamepad_detectDevices()) as often as it likes. I'd be interested to hear if you come up with any improvements to this pattern!

[Code_Assassin]

I made a tiny little API to POST/GRAB data in lovely JSON. Now I'm working on a server side script.