What Are You Working On? - August 2014

[Arxae]

Google tells me it was Okular, and the option is checked by default (according to wikipedia)

[MatheusMCardoso]

[QUOTE=Tamschi;45765569]Nice, but why would you check a Korean business website for exploits in the first place?[/QUOTE

Good question. I didn't. I just wrote the pfogram becauss it's fun. Fuck the spelling by the way. My phone is dumb.

[ECrownofFire]

[QUOTE=Map in a box;45762364]On the other hand you cant do much to put DRM on stuff like that.[/QUOTE]

In cryptography, Alice is trying to send a message to Bob without Eve eavesdropping.

With DRM, Bob is the same person as Eve.

[Xerios]

[IMG]http://i.imgur.com/SR5M6Tq.png[/IMG]

Released my game !

Give it a try :
[URL]http://facepunch.com/showthread.php?t=1419619[/URL]

[Edd]

[QUOTE=AntonioR;45757653]I made a little demonstration of Worms like destructible terrain made using Box2D chain loops and polygon clipping library called Clipper.

You can download it or read how it's done here:
[url]http://antonior-software.blogspot.com/p/destructible-terrain-worms.html[/url]
It's the first public test of my engine, so if you encounter any problems let me know.

[media]http://www.youtube.com/watch?v=Qkh8PMwOgQ4[/media][/QUOTE]

Oh my god thank you!

I'm working on 2d destructible worms like terrain myself at the moment and was thinking how to do the physics. Was thinking about implementing quad trees and doing the physics myself but I couldn't find any good tips on how to accomplish that. I guess I'll give Box2d another show. Great tips!

[Pelf]

[QUOTE=Xerios3;45769861][IMG]http://i.imgur.com/SR5M6Tq.png[/IMG]

Released my game !

Give it a try :
[URL]http://facepunch.com/showthread.php?t=1419619[/URL][/QUOTE]

Looks fun but it crashes on my phone :v:

[Xerios]

[QUOTE=Pelf;45770070]Looks fun but it crashes on my phone :v:[/QUOTE]

Could you provide more info about your phone? My google crash report doesn't show anything, maybe I could fix it somehow

Also, 4 hours since release. So far I'm not a millionaire yet... and 0 installs, things are going pretty smoothly !

[G4MB!T]

[QUOTE=Xerios3;45772017]Could you provide more info about your phone? My google crash report doesn't show anything, maybe I could fix it somehow

Also, 4 hours since release. So far I'm not a millionaire yet... and 0 installs, things are going pretty smoothly ![/QUOTE]

Welcome to indie game developer life :v:

[CapsAdmin]

I think I'm working on a GUI

animations, parenting and drag drop:
[vid]http://dl.dropboxusercontent.com/u/244444/ShareX/2014-08/2014-08-23_00-53-14.mp4[/vid]

framebuffers, scrolling and bringing stuff to front by using alpha from texture:
[vid]https://dl.dropboxusercontent.com/u/244444/ShareX/2014-08/2014-08-23_01-09-01.webm[/vid]

[vid]http://dl.dropboxusercontent.com/u/244444/ShareX/2014-08/2014-08-23_03-51-43.webm[/vid]

[Pelf]

[QUOTE=Xerios3;45772017]Could you provide more info about your phone? My google crash report doesn't show anything, maybe I could fix it somehow

Also, 4 hours since release. So far I'm not a millionaire yet... and 0 installs, things are going pretty smoothly ![/QUOTE]

Droid Razer M
Android 4.2.2
When I launch it, the screen rotates, is black for a second, then returns to the home screen

Dunno what else information you need

[rilez]

Hi guys. I'm new to C++, and programming in general. I've been lurking in this thread for a long time, watching you all make neat things. I know this is possibly the least interesting thing I could post, but I'm proud to know what all 15 lines of code do:

[img]http://i.imgur.com/futFEbc.pngg[/img]

[Xerios]

[QUOTE=Pelf;45772185]Droid Razer M
Android 4.2.2
When I launch it, the screen rotates, is black for a second, then returns to the home screen

Dunno what else information you need[/QUOTE]

Damn, I was hoping you'd have an older version of android and perhaps some api features didn't work ( even though I put everything to minimum ). Can't do much about it since it's all Uniteh three dee :(

[editline]23rd August 2014[/editline]

[QUOTE=rilez;45772486]Hi guys. I'm new to C++, and programming in general. I've been lurking in this thread for a long time, watching you all make neat things. I know this is possibly the least interesting thing I could post, but I'm proud to know what all 15 lines of code do:

[IMG]http://i.imgur.com/futFEbc.pngg[/IMG][/QUOTE]

Creation of my first hello world app was something I would never forget, I was so excited to see my own application with my own words, it was like magic back then

Welcome to the programming world !

[Gmod4ever]

So I arbitrarily decided to build a little utility that scans a directory (and its subdirectories) and creates a 3D map of it, mapping the last-time-modified versus the filesize versus the amount of files that fit into that bucket.

The idea is to run it on an entire file system (such as your C:/), and it'll map out your "file density," to help you determine what files you should delete when cleaning up your computer.

The program maps on a 2D grid, where the horizontal axis is time-modified, with the least recent time being at the left, and the most recent time being at the right. The vertical axis is filesize, with the largest files at the top, and the smallest files at the bottom. The third dimension is in color, much like a population density map: the brighter the color, the more files exist on your filesystem at that particular point. Therefore, when you run the program, the top-left corner of the map is the largest and least-recently-used files on your system: the ideal candidates for deletion when cleaning up your computer.

The program achieves this through a recursive, concurrent directory-crawl. It quantizes the last-modified and filesize of each file it scans, currently along 24-hour and 8MB boundaries, and then uses that vector as a key into a hashmap, with the value being a linked-list of files that fit within that boundary.

The program will support a method where you can select a region on the map, and it will display all of the files within that region (or at the very least, their filepaths).

I have the programmatic portion of the program working exceptionally well (and quick, in my opinion; the concurrent solution only took 50 seconds to scan my entire 250GB worth of files in C:/, as opposed to the original sequential solution that took over 160 seconds). I've only just started the graphical user interface.

So far, all I have to show for it is the function behind the graphical representation of file density: given a normalized percentage and an arbitrary-length array of colors, it will correctly generate an interpolated color between the two closest evenly-distributed points in that array.

Here is an image of it in action, which will later be used as my density scale. The top is the lowest density, and the bottom is the highest density. When it comes to making the actual 2D map, the scale will be flipped to fit the convention of the highest values being first (EG top, left, white is the absolute largest value possible).

[img]http://a.pomf.se/evwoyq.png[/img]

And the code used to generate:

[b]ColorInterpolator.java[/b]

package main;

import java.awt.Color;

public class ColorInterpolator {
	private ColorInterpolator() { }
	
	public static Color interpolate(float percent, Color[] colorPoints) {
		// Boundary cases
		if (percent < 0.0f) return colorPoints[0];
		if (percent > 1.0f) return colorPoints[colorPoints.length-1];
		
		// Find which two colors to interpolate between
		int interpStart = (int)(percent * (colorPoints.length - 1));
		
		Color start = colorPoints[interpStart+1];
		Color end = colorPoints[interpStart];
		
		// Inflate the percentage to be between these two points
		float step = 1.0f / (float)(colorPoints.length-1);
		float lower = interpStart * step;
		float upper = (interpStart+1) * step;
		
		float inflated = (percent - lower) / (upper - lower);
		
		// And interpolate between these two points
		Color result = new Color((start.getRed() * inflated + end.getRed() * (1.f - inflated))/255f,
				(start.getGreen() * inflated + end.getGreen() * (1.f - inflated))/255f,
				(start.getBlue() * inflated + end.getBlue() * (1.f - inflated))/255f);
		
		return result;
	}
}

[b]ResultImage.java[/b] (This will be expanded to draw several elements, and then merge them together. For now, it just draws the scale image. Likewise, its attributes will be relegated to variables that can be controlled during construction.)

package main;

import java.awt.Color;
import java.awt.Graphics;
import java.awt.image.BufferedImage;

public class ResultImage {
	private Color[] colorScale = new Color[]{ Color.BLACK, Color.BLUE, Color.YELLOW, Color.RED, Color.WHITE };
	
	public BufferedImage image;
	
	public ResultImage() {
		image = drawImageScale();
	}
	
	public BufferedImage drawImageScale() {
		int width = 32;
		int height = 400;
		
		int steps = 400;
		float fsteps = (float) steps;
		float scalestep = (float)height * (1 / fsteps);
		
		BufferedImage scale = new BufferedImage(width,height,BufferedImage.TYPE_INT_RGB);
		
		Graphics g = scale.getGraphics();
		g.setColor(Color.WHITE);
		g.fillRect(0, 0, width, height); // Fill out the background
		
		for (int i=0; i < steps; i++) {
			float p = i / fsteps;
			
			g.setColor(ColorInterpolator.interpolate(p, colorScale));
			g.fillRect(0, (int)(p * height), width, (int)scalestep+1);
		}
		
		return scale;
	}
}

[b]Core.java[/b]

package main;

import java.io.File;
import java.io.IOException;

import javax.imageio.ImageIO;

public class Core {
	public static void main(String[] args) {
		ResultImage result = new ResultImage();
		File out = new File("scale.png");
		
		try {
			ImageIO.write(result.image, "png", out);
		} catch (IOException e) {
			e.printStackTrace();
		}
	}
}

So yeah. That's what I've been working on the past few days. Haven't had anything to show until now, though.

[20k]

Well, this is 1 million triangles (all frontfacing) rendering at 60 fps, @1024x768. I really need to figure out how to adjust fragment sizes properly (ie how many blocks triangles are chopped up into) - too large, and small triangles have a lot of overdraw. Too small, and large triangles have far too many threads per triangle. Its a tradeoff that I need to fix

[IMG]https://dl.dropboxusercontent.com/u/9317774/1million.PNG[/IMG]

I also fixed little black holes appearing in everything. Turns out, I wasn't clamping texture coordinates to the border of tiles within a page (I have to allocate the textures manually in pages of a big 3D texture). This meant that texture reads were leaking every so slightly into other textures, and causing some pixels to appear black (as that was the only other thing the buffer contained at the time)

Edit:

Oh shit, i just fixed my fragment code. This is p cool :v:

Essentially:

Old code
[cpp] while(pcount <= fragment_size)
{
float x = ((pixel_along + pcount) % width) + min_max[0];
float y = ((pixel_along + pcount) / width) + min_max[2];

//do triangle business

pcount++;
}
[/cpp]

New code:

[cpp] while(pcount <= fragment_size)
{
float x = ((pixel_along + pcount) % width) + min_max[0];
float y = ((pixel_along + pcount) / width) + min_max[2];

if(y > min_max[3])
{
break;
}

//do triangle business

pcount++;
}
[/cpp]

Where min_max[3] is y's maximum bound. In too-large fragments, loads of unnecessary work gets done and y (drastically) exceeds the triangle's maximum extents. A simple 'If I'm past the end of the triangle, quit', has both taken performance from 16 -> 13ms in the million triangle case(!), but also allows me to set my fragment value to be very large (which is a better generalisation for all scenes). Basically this has fixed my entire triangle fragmentation problem (it was so simple, and i am so very angry/happy)

It turns out the extra checks of having to check if x is also out of bounds is totally not worth it, this way the check gets executed infrequently and pays itself off dramatically rather than forcing the check every loop iteration

[lumaio]

All these fancy 3D OpenGL things and all I have are some flat spinning triangles that I wrote in Go.
[VID]http://a.pomf.se/eyljpi.mp4[/VID]

[editline]23rd August 2014[/editline]

ShareX slowed my programming drastically. It dropped from 60 fps to 30. Annoying.

[AwesomeX]

Although I'm not working on a ton at the moment. I just wanted to say how awesome it is to be connected to a bunch of folks who are creating amazing stuff everyday. It's really incredible seeing what some of you can do.

I hope one day I can achieve what some of you are doing.

[Geel]

[QUOTE=sarge997;45773196]Although I'm not working on a ton at the moment. I just wanted to say how awesome it is to be connected to a bunch of folks who are creating amazing stuff everyday. It's really incredible seeing what some of you can do.

I hope one day I can achieve what some of you are doing.[/QUOTE]

The secret is we all cheat

Take layla for instance. All he does is take pictures of source maps.

[20k]

[QUOTE=Lumaio;45773090]All these fancy 3D OpenGL things and all I have are some flat spinning triangles that I wrote in Go.

[editline]23rd August 2014[/editline]

ShareX slowed my programming drastically. It dropped from 60 fps to 30. Annoying.[/QUOTE]

Blimey, I fooled someone thinks my renderer is OpenGL. Life's work = complete

For reference, its a near pure OpenCL implementation of a deferred 3D renderer

Obligatory sponza atrium:

[IMG]https://dl.dropboxusercontent.com/u/9317774/ifooledsomeone%21.PNG[/IMG]

The main thing I'm lacking is dynamic shadows at the moment, which is a bit annoying. I should also probably implement radius lights for more dramatic deferred-ness, but that would be vaguely boring

[Map in a box]

[QUOTE=rilez;45772486]Hi guys. I'm new to C++, and programming in general. I've been lurking in this thread for a long time, watching you all make neat things. I know this is possibly the least interesting thing I could post, but I'm proud to know what all 15 lines of code do:

[img]http://i.imgur.com/futFEbc.pngg[/img][/QUOTE]

Who's to say that's not neat? Also, you should use getchar instead of cin >> ch IMO. Printf is awesome, too. I'm always a little bit secretly giddy when some lurker joins in the sweatshop of WAYWO.

[DrLuke]

[QUOTE=Xerios3;45772733]
Creation of my first hello world app was something I would never forget, I was so excited to see my own application with my own words, it was like magic back then

Welcome to the programming world ![/QUOTE]

I'll never forget my first blinking LED.

[James0roll]

Working on procedural texturing. It seems like a much more practical technique compared to storing a bunch of textures. Rather, use an algorithm to generate the material. By doing this, it saves memory, and could eliminate repeating textures. I haven't seen [b]anybody[/b] use such a method before, but it seems like such a practical idea.

As seen in the video, rendering a texture 2048 by 2048 takes roughly 5 seconds, which includes the normal mapping as well. Add another texture for texture blending, and it might take 10 seconds. Now add the textures for the rest of your game (sand, dirt, grass, water, wood, brick, tiles), and it might take 1+ minutes, but at least you have organic textures without the hassle of memory limits and trying to make the textures yourself.

I am rendering (fine-grained) sand, which is really just a bit of noise, but I will achieve bricks, grass, and etc..

I hope to add:
- Texture blending
- Loading screens
- [b]More materials[/b]

Once I feel satisfied with this, I may advance to applying it into a game. I was thinking virtual reality.

[video=youtube;ZWii1ArLsYI]http://www.youtube.com/watch?v=ZWii1ArLsYI[/video]

I am doing this in Unity, but the whole concept isn't necessarily Unity restricted, so I felt it deserved to be here.

[WTF Nuke]

[QUOTE=Map in a box;45773886]Who's to say that's not neat? Also, you should use getchar instead of cin >> ch IMO. Printf is awesome, too. I'm always a little bit secretly giddy when some lurker joins in the sweatshop of WAYWO.[/QUOTE]

Correct me if I'm wrong, but I thought getchar is deprecated and a carryover from C much the same as Printf (the C part).

[G4MB!T]

[QUOTE=WTF Nuke;45774574]Correct me if I'm wrong, but I thought getchar is deprecated and a carryover from C much the same as Printf (the C part).[/QUOTE]

I dont think its deprecated. Its in there for compatibility reasons (has C ever actually deprecated anything?), but you are right that it is from C.

You can use:

std::cin.get();

Which, unlike things like system("pause"), will work on multiple (if not all) platforms.

[lumaio]

[QUOTE=WTF Nuke;45774574]Correct me if I'm wrong, but I thought getchar is deprecated and a carryover from C much the same as Printf (the C part).[/QUOTE]

You are correct. But they are still in every way superior.

[Darwin226]

[QUOTE=Lumaio;45775028]You are correct. But they are still in every way superior.[/QUOTE]

Except the ways that are important.

[WTF Nuke]

I don't really understand how they are superior. Sure it's easier to output things with printf due to it automatically concatenating your string with additional arguments, but it has a lot less type safety which I think is very important in modern C++. I also don't like mixing C and C++, but I guess that's personal preference. I find it best to stick to C++ functions and very rarely rely on C calls, as there is almost always a C++ alternative.

[ECrownofFire]

[QUOTE=WTF Nuke;45774574]Correct me if I'm wrong, but I thought getchar is deprecated and a carryover from C much the same as Printf (the C part).[/QUOTE]

Nothing from C is deprecated. Except gets(), but that's also deprecated in C.

[WTF Nuke]

That's the one I'm thinking of.

[ECrownofFire]

[QUOTE=WTF Nuke;45775310]I don't really understand how they are superior. Sure it's easier to output things with printf due to it automatically concatenating your string with additional arguments, but it has a lot less type safety which I think is very important in modern C++. I also don't like mixing C and C++, but I guess that's personal preference. I find it best to stick to C++ functions and very rarely rely on C calls, as there is almost always a C++ alternative.[/QUOTE]

Most of the arguments I've seen surrounding cout vs printf are based around how awkward formatting is.

[cpp]
std::cout << std::hex << foo << std::endl;

printf("%x\n", foo);
[/cpp]

[Natrox]

I just prefer to use printf most of the time; it's much more clear to me and doesn't result in long/multiple lines of std::cout stream stuff.

Also for decimal precision, I prefer the printf version;

std::cout << std::setprecision( 5 ) << val;

printf( "%.5f", val );