next »

[Rob Lach]

I'm almost done mimicing the acceleration map of my volvo using info I found online on the volvo gear ratios.

[bateleur]

Hacking my way through the last few KMScript bugs.

Then I'll write a script to make me some lunch.

[_Tommo_]

Hacking my way through the last few KMScript bugs.

Then I'll write a script to make me some lunch.

Just... why not just separate those two layers (it's not unusual to have them, btw) in C++ without the hassle of writing a whole new language?
If it can be done in a "scripting" language it can be definitely done in the "host" language, after all the term scripting comes from this: a script is a sequence of commands that drives the "real" code.
Are you sure you couldn't pull off the whole "now push is a generic action on a thing" natively?

Btw: yeah I know it's fun  

[bateleur]

Just... why not just separate those two layers (it's not unusual to have them, btw) in C++ without the hassle of writing a whole new language?

Assuming that by "C++" you mean "what the rest of the game's written in" (UnityScript, as it happens), there's a combination of reasons:

1) Awkward to support code generated at runtime.
2) Concerns over maintaining sandbox integrity.
3) Complexity of expressing conceptually simple ideas.

Are you sure you couldn't pull off the whole "now push is a generic action on a thing" natively?

Clearly I could, but I'm not sure what the upside would be of doing it that way.

Btw: yeah I know it's fun  

Yep, that too!

[_Tommo_]

when you talked about scripting I just assumed you came from C++ 

In C++, the upside is to retain the expressiveness and the power of native code... you don't have to write a wrapper and you don't have to fight with scripting limitations; you just need to be clear on what belongs to which layer (and using libs and/or namespaces it is easy).

But, I don't know US, so I take that you know what you do 

[bateleur]

But, I don't know US, so I take that you know what you do 

No, I think your comments still apply.

You wouldn't seriously compile C++ at runtime... or would you?

[rivon]

But, I don't know US, so I take that you know what you do 

No, I think your comments still apply.

You wouldn't seriously compile C++ at runtime... or would you?

http://bellard.org/tcc/

[Nix]

z-buffer output (this is drawn to a single component image that is attached to the g-buffer FBO and then drawn back onto the screen as a full-screen quad): http://dl.dropbox.com/u/10832827/z-buffer.ogv for my wip deferred shading renderer

[_Tommo_]

No, I think your comments still apply.

You wouldn't seriously compile C++ at runtime... or would you?

Ok, compilation times are definitely a downside 
I never tried to compile code at runtime, I admit it is cool

@rivon: how would you move to use tcc to replace existing code segments at runtime? I suspect that it involves crazy amounts of crazyness.

[ThemsAllTook]

This thread used to be as active as the happy and grumpy programmer rooms, but I had to go all the way to page 35 to dig it up! Dunno if it would have been better to start a new one, but let's see if this one has any life left...

I've started on a remake of one of my old games, and I'm currently trying to figure out how I want the visuals to work. I've never entirely felt good about using bitmaps (too resolution dependent), and I don't want to venture into the territory of 3D models at the moment, so I'm thinking of trying out SVGs. I've found a nice-looking SVG parser, so assuming it's up to the task and I don't have to write my own, that just leaves writing a renderer (and creating the graphic assets, of course). Lots of interesting territory to explore here... Can I render an arbitrary SVG with OpenGL and anti-alias it nicely? How do I handle animation? I guess most people who want to use this kind of vector graphics in games do it with Flash... I should probably take a look at how it handles animation before inventing a system of my own for it. Any experiences or tips to share?

[InfiniteStateMachine]

Fixing a program I was given written in Python. I don't really know python so I've been studying up on idomatic python so I don't do anything too bad.


Unfortunately this code seems to follow some strange conventions.

TBH I'd prefer to rewrite the program in C but when in rome...

[Twinklebear]

I'm working on migrating my old system of hard-coded Button classes that can only call one function with varying parameters (to change scenes), to be their base class that's managed by a Lua script which will call some function to execute the button effect when it's clicked. I've got that together but now need to create a lookup/dispatch system for the engine that can send the function call to the right place, which may be a bit tricky, especially with not knowing what parameters will be passed/how many and so on, because the desired function could be a user-defined function in some other entity/state's lua script.

I think it'll be a bit tricky to get sorted, but it's definitely an interesting challenge.

[Grungi_Ankhfire]

I'm assembling parts of a voxel editor in Panda3D (so it is Python code). I did the camera controls this morning (orbiting, panning, zooming), I have been researching automatic mesh generation algorithms, and will start placing and removing voxels tonight!

[R.D.]

Doing some procedural generated music that sounds good. Shit lot of music theory here... I combine it with my self designed gui in TWL.

[moi]

right now I have started writing an integrated tool, that I'm not sure I'll ever use, but by programming it I hope to understand a few truths about asset creation (my weakness).
It is an all in one program that should help create character sprites with an old-school arcadey style.
There are 3 parts, a part to design limbs (hierarchically linked , and hopefully will save anims to a skeletal format in the future), a part to paint on the limb  (each limb is like a sprite) using very oldschool Amiga techniques and color ramps, and a part to create the color scales