Quakeboy'z Dev Arena

Lets learn pixel shaders - Chap 2 - Basics

2011-12-26T19:21:00.000+05:30

Important - To move forward open another tab in your browser and open the URL http://www.iquilezles.org/apps/shadertoy/, I will refer to this tab simply as 'Shader Toy'

Chapter 2 - Basics of pixel shader with 'Shader toy'

Let us write some code, switch to the shader toy and notice that the "Deform" shader is already loaded and is running by default. On the right side bottom you see the source code, and on the left is the rendered effect. Take some time to notice the various details/controls shown to you like

1) fps, time, play/pause button, restart button on top of render area (place where the effect is played).

2) Dimensions panel - Width and height of the render area on top it. You can change it if you want.

3) The textures passed to the pixel shader are given right to the width/height box. You can pass up to 4 textures and use them in the pixel shader. The texture used in deform shader is given in the box 'unit 0'. Copy paste that URL path in a new tab and check the texture for yourself, or you can just click this link.

4) Also you can check the other shaders by going through presets drop down list. Don't forget to click 'Load'. There are some cool effects which you have no idea how to create, but at the end of this series you will learn to make most of them.

5) Most important thing we need to notice is the compile button. It is given above the source code area along with 'help' button.

n00b note - Every time we write new code, we need to press the compile button to see new source code taking effect.

6) Now let us write some code. We will begin with a blank program. Copy paste the following code in the code window.

#ifdef GL_ES
precision highp float;
#endif
uniform float time;
uniform vec2 resolution;
uniform vec4 mouse;
uniform sampler2D tex0;
uniform sampler2D tex1;
void main(void)
{
}

--- Now hit the compile button and see the output on the left side ---

7) Beautiful isn't it !
It will be a blank screen. Basically for now we have not written any code. Being a programmer I'm sure you can notice that main function is empty. If you don't know what main() function is, then stop reading right away, don't waste your time. Grab a good C or C++ book or tutorial.

8) Now let us write our first line. If you recap I said - "pixel shaders are programs that are run for every pixel on the screen and they can do one thing, set the pixel color". So let us do that now. Inside the void main function block write the following line

gl_FragColor = vec4(0.0, 0.0, 0.0, 1.0);

gl_FragColor is a built-in variable through which we can write the pixel color. vec4 is a custom data type just like int, float, bool etc. But vec4 is a vector of 4 float numbers. The numbers are the colors - R, G, B and A. So here we say R, G, B are zero and Alpha is one which works out to black color. Hit compile button and make sure it is not paused on the left side, you will see the screen turning black. Now try replacing that line with the following.

gl_FragColor = vec4(1.0, 1.0, 1.0, 1.0);

Click compile and run again, you get the idea? As the same program is running for each pixel on the rendering area, and what we are doing is just setting it to white, the whole screen looks white.

Now how do we make things interesting? Wait for next chapter :)

Lets learn pixel shaders! - New Series - Chapter 1

2011-12-25T16:09:00.001+05:30

This is the beginning of a series of tutorials, about pixel shaders (aka fragment shaders) with WebGL. I will start by assuming you are someone who has "absolutely no experience in graphics programming", but you should know what pixels are, what their components are. And optionally if you have basic 3d texturing knowledge, it will help a lot, but again not compulsory.

Without more blah, blah, let us start learning fellow programmers!

Thanks to http://www.iquilezles.org/apps/shadertoy/

without which I would have never even started this! All our tutorials will be explain with reference to that web app. You will need one of the latest browser to run it.

Chapter 1 - Pre-requisite (for the n00b graphics programmer)

Pixels (skip if you know)

Everything you see on your monitor is made of pixels. If you zoom into your computer monitor, basically it is made of basic blocks (usually square shaped) called pixels. But as they are very small and closely packed, you are not able to notice them. Every monitor doesn't have the same number of pixels.

Resolution (skip if you know)

It is (number of pixels lengthwise) x (number of pixels widthwise). For eg. iPhone 3GS has a resolution of 320 x 480 when you hold the phone in portrait mode. iPhone 4 / 4S has 960 x 640 as their resolution. A full HD (High Definition) monitor has 1920 x 1080 pixels. More the resolution, more is the image detail/clarity. Will you be able to show more details with a 10x10 grid of dots or a 100x100 grid of dots? Think about it.

Pixel components (skip if you know)

Each pixel can do only one thing - "They can display a color". A pixel has three basic components - R, G and B. They are the primary colors, and combining them in various ways, we can make thousands (or millions, in latest monitors) of other colors. Usually each color is represented with one byte (8 bits, so value range is 0 - 255). So put together for each pixel you will have total of 3 x 8 bits = 24 bits or 3 bytes of color information.

For eg.
black (R = 0, G = 0, B = 0)
white (R = 255, G = 255, B = 255)

Introducing Alpha (skip if you know)

Apart from R, G and B, there is another component for a pixel, its called Alpha. Alpha = Transparency. Often pixels can be overwritten on top of each other, when that happens they are "mixed" according to their alpha levels. So R, G, B and A are the components of a pixel, which takes 32 bits per pixel. Understand there are more ways to represent a pixel, its not always 32 bits, but it is the most advanced pixel format, and enables an individual pixel to represent the max number of colours.

So, what are pixel shaders ?

Now that you know what pixels are, pixel shaders are nothing but programs, very similar to a C program. Next paragraph is very important.

A pixel shader is run for every pixel, before getting displayed on the screen. So let us say we have 1920 x 1080 pixels for a full screen game, which is a total of 2073600 pixels, which means a pixel shader is run that many times! But that is okay because they are all run on the GPU, instead of the CPU (Graphics processing unit, a special hardware dedicated to graphics, present nowadays on almost all computers).

Now that we know the basics, let us get started with
1) what can be done in a pixel shader,
2) how it can be done and
3) what information we have access to in the next chapter !

Any doubts are welcome in the comments section.

Common mistakes done by every (game) programmer

2011-07-22T12:23:00.000+05:30

This is a re-blog of a recent article in programming section of Gamasutra. It was titled "Programming Sins - Common Errors From Down In The Trenches". I would like to summarize it, but I definitely recommend you to read and start adding some points to your coding standards document.

Here is what I thought are the main points from that article

Coding standards for Games.

1. Always code for NULL cases when loading resources and using it. Expect NULL data where ever possible and code.
2. When you write functions DO NOT use ambiguous variable names, function names and parameter names.
3. Do NOT use abbreviations when naming variables/ functions/ parameters.
4. Validate pointers at least once before accessing them.
5. Validate data (clip them to valid ranges etc.)
6. Have a default state (like a default texture)
7. Consider Memory fragmentation when planning your code.
8. AVOID copy pasting code, try to always make them into functions when reusing.

You should definitely read the original article here though, as it convinces and gives excellent real world examples of why you should follow the above points.

XCode 4 icon

2011-04-12T16:25:00.000+05:30

Like me if you have installed both XCode3 and XCode4, and you wanna have both in the dock, you will find having the same icon is confusing, so here is a custom icon which you can use.

Right click XCode.app and click "Show Package Contents" and after that navigate to "Contents/Resources" and replace the existing Xcode.icns with this one.

Download Custom Xcode.icns file.

Download Android SDK standalone for offline installation

2011-02-22T13:02:00.000+05:30

Update :- Ice Cream Sandwich links (4.0) updated - thanks Vikram!

How to install Android SDK without internet connection

I searched all over the internet and found no posts like this, hence I'm making one hoping it would be helpful for a lot of people.

The magic URL used to be - http://dl-ssl.google.com/android/repository/repository.xml (Outdated)
That is the XML file from which the URL for downloading the SDK packages are obtained.

Update :- The previous URL is now invalid, the new URL is given below
https://dl-ssl.google.com/android/repository/repository-5.xml

For e.g. if you want to download Android SDK for version 4.0.3 for all platforms, you could look up that XML file. You will find a block under tag SDK 4.0.3 like this

<sdk:archive arch="any" os="any">

<sdk:size>44758833</sdk:size>
<sdk:checksum type="sha1">f2aa75133c29916b0b0b984974c2d5e46cb09e9c</sdk:checksum>
<sdk:url>android-15_r01.zip</sdk:url>
</sdk:archive>

So the URL will be is http://dl-ssl.google.com/android/repository/android-15_r01.zip

If all the above sounds too complex

You can directly download from the below links. If this page does turn out to be useful, then I might update it regularly. Do leave a comment to let me know if it has helped you.

Android SDK Platform 4.0.3 (*** NEW ***)
All Platforms

Android SDK Platform 4.0
All Platforms

Android SDK Platform 3.2 revision 1
All Platforms

Android SDK Platform 3.1 revision 3
All Platforms

Android SDK Platform 3.0, revision 2
All Platforms

Android SDK Platform 2.3.3_r2
All Platforms

Android SDK Platform 2.3.1_r2
All Platforms

Android SDK Platform 2.2_r3
All Platforms

--- OLDER LINKS (Still working) ---

Android SDK Platform Honeycomb Preview, revision 1
All Platforms

Android SDK Platform 2.3.3_r1 (Offline complete download) - Approx 80 MB
All Platforms

Android SDK Platform 2.3_r1 (Offline complete download) - Approx 80 MB
All Platforms

Android SDK Platform 2.2_r2 (Offline complete download) - Approx 80 MB
Windows
Mac
Linux

Android SDK Platform 2.1_r1 (Offline complete download) - Approx 80 MB
Windows
Mac
Linux

Android SDK Platform 2.0.1_r1 (Offline complete download) - Approx 80 MB
Windows
Mac
Linux

Android SDK Platform 2.0, revision 1 (Offline complete download) - Approx 80 MB
Windows
Mac
Linux

Android SDK Platform 1.6_r2 (Offline complete download) - Approx 63 MB
Windows
Mac
Linux

Android SDK Platform 1.5_r3 (Offline complete download) - Approx 54 MB
Windows
Mac
Linux

Android SDK Docs for Android API 7, revision 1 - Approx 48 MB
All Platforms

Android SDK Tools, revision 5 (16 to 23 MB)
Windows
Mac
Linux

Pointer not declared even when declared !

2011-02-17T18:12:00.000+05:30

A Strange problem encountered yesterday which left us puzzled for a few mins :) (Check image below)

So now, we tried everything possible. Until at last when we saw that there was a backslash after "Plist" in the comment line. As backslash is an escape sequence, it made the whole next line invalid, more like a comment.

Once we deleted that backslash, things were back to normal. But hey, aren't comments supposed to suppress anything in them including escape characters

How to ignore build directory in SVN from command line

2011-02-17T17:58:00.000+05:30

This works for Windows/Mac/Linux, but its tested on Mac. On Windows, you can do it easily with Free TortoiseSVN Shell context menu, so this method will be useful on Mac/Linux.

Lets suppose your folder structure is
SVNfolder->trunk->development->Projectfolder->build

Now open terminal and cd to development. (Now if you 'ls' it wil' list Projectfolder)
type svn propset svn:ignore "build" Projectfolder/ and press Enter.

and that is it, your build folder will be ignored.

Simflex Image Gallery - Full Free Flex Image Gallery

2009-12-20T18:17:00.000+05:30

I am releasing my source code for a full fledged Flex 3 image gallery along with a PHP based Minimal Image file manager page with upload, delete and automatic thumbnail generation features. I had made this while learning Flex, but it is very usable in many projects.

To see it in action, download, extract and put the bin-debug folder under your web server. There is a folder called "Admin" inside "bin-debug" folder which has the PHP part responsible image upload, XML file generation and auto thumbnail code. To manage the images, use filelist.php in admin folder. Tested with JPG and PNG images.

Download Source code - 8.78 MB (along with sample images)
Please drop me a line if you are using this anywhere.

3d Tunnel OpenGL Source code - iPhone OpenGL ES compatible !

2009-12-20T12:25:00.000+05:30

I am releasing source code of the iPhone OpenGL ES compatible Never-Ending Tunnel Source Code. The code is done using SDL on Windows only. But the OpenGL elements use only Triangle strips and all OpenGL ES compatible functions and porting to iPhone is a cakewalk.

I have searched on the web and never found a ES compatible version which is easily portable, so here we go enjoy. I have intentionally left a small glitch in Texture Coordinate generation, which actually looks pretty according to me. If you need I can help you to fix is easily.

Some important notes :-

No Models present - The Tunnel is a torus which is completely procedurally generated.. so the torus properties can be changed easily.
The Texture generation part is platform dependent. The files RTexture.h and RTexture.cpp is almost completely copy-paste reusable :D
DO NOT REMOVE - #define QUAKEBOY_IS_AWESOME (!)

Download Source code (along with SDL Library and Binaries included - 1.36 MB)

Andre Michelle's Tile based Scrolling ported to AS 3

2009-09-26T16:45:00.001+05:30

Update:- I have found something that has to be added to the existing code. The SWF embedded here runs only @ around 60 FPS on a Safari or Firefox on a Mac. I will work on a more consistent render loop code adding some timer calculation elements soon.

I have ported the Tile based scrolling in Flash code by Andre Michelle to use Actionscript 3.0
Main challenge I faced was the tearing issue which was removed after I used Timer Event and updateAfterEvent() method. Thanks to 8bitRocket site for the tip.

I am getting a blazing 160 FPS in browser.. OMG Flash rocks ! If anyone wants me to provide comments and documentation for the code. I will happily do it on request.

Click here to get the full source code.

P.S. If its slow, its because you might be having the static noise swf playing below (scroll page down). View on the post page by click the title to ensure it plays smoothly.

Creating Static TV Noise in Flash using AS3

2009-09-16T00:25:00.000+05:30

Creating Fast and Smooth Static TV Noise in Flash using AS3 was a simpler task than I expected it. First attempt was using for loops to set every pixel and was not very fast. A Simple browse through the API docs for BitmapData class gave me the 'noise' method

bgdata.noise(Math.random() * 1000, 0, 255, 7, true);
did the trick :)

Complete code below

package {

import flash.display.Sprite;

import flash.events.Event;

import flash.display.Graphics;

import flash.display.Bitmap;

import flash.display.BitmapData;

import flash.geom.Point;

public class StaticNoise extends Sprite {

const ww = 480;

const hh = 272;

var bgdata:BitmapData = new BitmapData(ww, hh, false, 0xffffff);

var bg:Bitmap = new Bitmap(bgdata);

var zeroPt:Point = new Point(0,0);

public function StaticNoise() {

addChild(bg);

addEventListener(Event.ENTER_FRAME, drawBaby);

}

function drawBaby(e:Event) {

bgdata.lock();

bgdata.noise(Math.random() * 1000, 0, 255, 7, true);

bgdata.unlock();

}

You can modify,use this code where ever you want, without giving credits. But please just let me know in the comments if it helped you.

XCode code comes to center of screen problem

2009-07-10T11:28:00.000+05:30

Today I had a strange error with code alignment when using XCode. After I pressed ENTER key, the next line begins from the center of screen. Each time I have to manually press the backspace and it was annoying.

Solution

Select All
Right Click/Ctrl Click
Select "Re-Indent Selection"
Now examine the full file by scrolling from top to bottom.
If the code starts aligning itself strangely from a particular line, the problem is with the previous line. You might have probably missed out a bracket (which was the case with me) or something else too..

Hope XCode developers @ Apple fix this in the next release.

iPhone OpenGL ES Extensions List

2009-07-09T13:09:00.000+05:30

On the Simulator
iPhone 3.0 SDK with Simulator 3.0

GL_OES_compressed_paletted_texture
GL_OES_draw_texture
GL_OES_mapbuffer
GL_OES_matrix_palette
GL_OES_point_size_array
GL_OES_read_format
GL_OES_stencil8
GL_OES_texture_mirrored_repeat
GL_OES_point_sprite
GL_OES_framebuffer_object
GL_OES_blend_subtract
GL_EXT_texture_lod_bias
GL_EXT_texture_filter_anisotropic
GL_OES_depth24
GL_OES_rgb8_rgba8
GL_IMG_read_format
GL_IMG_texture_compression_pvrtc
GL_IMG_texture_format_BGRA8888
GL_APPLE_texture_2D_limited_npot

On the Device
iPhone 3G with OS 3.0

GL_EXT_texture_filter_anisotropic
GL_EXT_texture_lod_bias
GL_IMG_read_format
GL_IMG_texture_compression_pvrtc
GL_IMG_texture_format_BGRA8888
GL_OES_blend_subtract
GL_OES_compressed_paletted_texture
GL_OES_depth24
GL_OES_draw_texture
GL_OES_framebuffer_object
GL_OES_mapbuffer
GL_OES_matrix_palette
GL_OES_point_size_array
GL_OES_point_sprite
GL_OES_read_format
GL_OES_rgb8_rgba8
GL_OES_texture_mirrored_repeat

On the Device
iPod Touch 2nd Gen with OS 3.0

GL_EXT_texture_filter_anisotropic
GL_EXT_texture_lod_bias
GL_IMG_read_format
GL_IMG_texture_compression_pvrtc
GL_IMG_texture_format_BGRA8888
GL_OES_blend_subtract
GL_OES_compressed_paletted_texture
GL_OES_depth24
GL_OES_draw_texture
GL_OES_framebuffer_object
GL_OES_mapbuffer
GL_OES_matrix_palette
GL_OES_point_size_array
GL_OES_point_sprite
GL_OES_read_format
GL_OES_rgb8_rgba8
GL_OES_texture_mirrored_repeat

Thanks to Jeff Lamarche

Code to get all OpenGL ES extensions

NSString *extensionString = [NSString stringWithUTF8String:(char*)glGetString(GL_EXTENSIONS)];
NSArray *extensions = [extensionString componentsSeparatedByCharactersInSet:[NSCharacterSet whitespaceCharacterSet]];

for (NSString *oneExtension in extensions)
NSLog(oneExtension);

------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Conclusion:- There is no difference in OpenGL ES Extensions between an iPhone and iPod touch versions used in tests above. But in Simulator, it has everything supported by Device + Extra Extensions too.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Android keyCodes list

2009-04-14T12:16:00.000+05:30

a - z
29 - 54

"0" - "9"
7 - 16

BACK BUTTON - 4
MENU BUTTON - 82

UP, DOWN, LEFT, RIGHT
19, 20, 21, 22

SELECT (MIDDLE) BUTTON - 23

SPACE - 62
SHIFT - 59
ENTER - 66
BACKSPACE - 67

How to take Screenshot in Android Emulator with Eclipse

2009-04-06T12:27:00.000+05:30

I used to use PrintScreen > MSPAINT > Ctrl+V combination to capture screen of Android Emulator. Recently discovered that we can get a neat screen shot inside Eclipse using the Devices View.

If the Devices panel is not visible, on the menu bar click Window > Show View > Other. In the newly opened dialog box, under Android category, select Devices.

Now on the Devices panel, Click the button as shown in the Image and your screenshot is ready.

tags:- android screen capture, android screenshot, android emulator screenshot.

OpenGL texture upside down in Android ?

2009-02-11T16:12:00.000+05:30

Yes....

The texture co-ordinates in Android for OpenGL ES has been to be calculated with upper left as origin 0,0 as opposed to the traditional OpenGL where bottom left is the origin (0,0)

Remember upper left/ top left is the origin (0,0)

How to load a texture in Android OpenGL ES

2009-02-09T19:09:00.001+05:30

Here is how you can easily load a texture from any image format - PNG, BMP, JPG etc.. in OpenGL ES in Android.

**Before you begin the code, you need to import your images into your project's /res/drawable/ folder. Just right click the project in eclipse and select import to do that.**

//First setup the integer array to hold texture numbers which OpenGL generates
int texture[] = new int[1];

//Generate and bind to the texture (gl is my GL10 object)
gl.glGenTextures(1, texture, 0);
gl.glBindTexture(GL10.GL_TEXTURE_2D, texture[0]);

//Setup the texture co-ordinates
float texCoords[] = {
0.0f, 0.0f,
1.0f, 0.0f,.......};
FloatBuffer texcoords = FloatBuffer.wrap(texCoords);
gl.glTexCoordPointer(2, GL10.GL_FLOAT, 0, texcoords);
Bitmap wood = BitmapFactory.decodeResource(this.getContext().getResources(), R.drawable.myimagefilenamewithoutextension);

//Setup optional texture parameters
gl.glTexParameterf(GL10.GL_TEXTURE_2D, GL10.GL_TEXTURE_MIN_FILTER, GL10.GL_LINEAR);
gl.glTexParameterf(GL10.GL_TEXTURE_2D, GL10.GL_TEXTURE_MAG_FILTER, GL10.GL_LINEAR);

//Set the texture image
GLUtils.texImage2D(GL10.GL_TEXTURE_2D, 0, wood, 0);

//Enable texture related flags (Important)
gl.glEnable(GL10.GL_TEXTURE_2D);
gl.glEnableClientState(GL10.GL_TEXTURE_COORD_ARRAY);

Hope that helped you.. please let me know any problems

Turning on Depth Buffer makes everything disappear !

2009-02-05T15:18:00.000+05:30

I finally got to use textures on Android using OpenGL ES, and I found that its turn I started using the Depth buffer for some proper rendering. The moment I uncommented

gl.glEnable(GL10.GL_DEPTH_BUFFER);

The object disappeared and became invisible. Not even a pixel of the object was visible. Then a colleague of mine gave me a hint about clearing buffers !

Uffff ! I corrected the line
"gl.glClear(GL10.GL_COLOR_BUFFER_BIT);"
--to--
"gl.glClear(GL10.GL_COLOR_BUFFER_BIT | GL10.GL_DEPTH_BUFFER_BIT);"

and everything was fine again !

Funny Resume

2009-02-05T14:46:00.000+05:30

I fell on http://uk.geocities.com/sloppyturds/cv.html from one of gamedev.net forum posts. Then I found a CV which was interesting and funny (at least for me !)

zeds cv

i am 29 years old kiwi who born in aoetearoa on the look out for game programming work.

talents/deeds/wishes

ive written 2 operas + more than 400 songs

i can drink lots of piss without getting pissed

can program c, c++, pascal, basic, z80 assembly, cobol(the only one ive been offically taught). some though i havent done for 'n years.

basic knowledge of win32 + linux + opengl + directx ( though i will not use the fascist api ).

i speak 3 languages (fluent in none - not even my mothertongune)

well travelled (7x round the world)

i need classical music to work esp wagner

one of the strongest ppl on the planet, physically and mentally

can work 80 hour weeks

havent washed my hair this century, first person to fart this century (unofficial).

plays a resonable accoustic guitar

willing to relocate to + live in, in order of preference ( no other countries!!! )

- eastern europe - south germany - iceland - china - rest of europe - se asia - israel - south america - india - oz ( except queensland - too many wankers )

dont delay, hurry up and grab this ass-et for your programming team.

Using M_PI gives error even after including math.h ?

2009-01-23T16:24:00.000+05:30

If you look at the math.h header file, the PI math constant M_PI and other values will be defined at the end of the file and just above that there will be a conditional #define pre-processor statement.

Solution :- Just type in #define _USE_MATH_DEFINES ~~after~~ before including math.h

#define _USE_MATH_DEFINES
#include <math.h>

class type re-definition error - Solution !

2009-01-19T12:24:00.000+05:30

I am writing this to help other people avoid spending too much time on this silly error.
I spent 45 mins after which I realized -

I missed out #pragma once on the .h file of the class declaration.

If the above solution didn't help you, you might search for another identity (class, function , variables) with the same name as the class name which gives the error.

TCHAR array, std::string or std::wstring - none of them !

2008-12-18T17:59:00.001+05:30

Having confusion about how to make flexible use of strings in your C++ application. Here is the solution I got after googling. I love the way C++ macros help in maintaining code well

You don't have to use TCHAR, std::string or std::wstring in your application directly and make it rigid. Make use of the code below and then use the typedef - String or tstring. Now your strings in your C++ application becomes very flexible.

Converting 'TChar' to 'String' in C++ - GameDev.Net Discussion Forums

TCHAR is just a typedef that depending on your compilation configuration (whether you have #define'd UNICODE or not) defaults to char or wchar. Note that you don't necessarily need to #define UNICODE as part of your code. You may also pass it to your target compiler via some preprocessor definition argument, such as /D in microsoft's visual C++ compiler.

Standard Template Library on the other hand supports both ASCII (with std::string) and wide character sets (with std::wstring). All you need to do is to typedef String as either std::string or std::wstring depending on your compilation configuration, if flexibility is what you're aiming for. Something like this might prove useful:

#ifndef UNICODE
typedef std::string String
#else
typedef std::wstring String
#endif

converting std::string to TCHAR

You could do a compile time version of string like this:

#include

namespace std {

#if defined _UNICODE || defined UNICODE

typedef wstring tstring;

#else

typedef string tstring;

#endif

}

Then you could use tstring instead of string and it would depend on the compile options.

Tom

SVN Very quick start tutorial/guide

2008-10-10T18:06:00.001+05:30

I recently setup a SVN server successfully on Windows XP and learnt to use a Visual Studio 2005/2008 Plugin called Ankh SVN (Previous versions of this plugin seem to be buggy, but the latest one kicks ass !).

Windows Servers are available from here. I used this one.

Steps:-

Download and install the latest SVN Server.
I set up only the SVNServe standalone server and not the Apache addon module. (I read a comparision study revealing SVNServe is faster/better than Apache module)
Then select the repository directory. (Here I will use - g:\repos)
Complete the installation.
Open CMD
Type net start svnserve. (You have to start server once only, on next reboot it will start automatically)
Change directory to g:\repos by cd command.
Type - svnadmin create myrepos. Repository is created.
Open the svnserve.conf file under the conf folder in the created repository (g:\repos\myrepos).
Uncomment the line password-db = passwd
Open passwd file. Under [users] block remove the commented example users (Harry etc). Create users in the format <username = password>.
Now download & install AnkhSVN for Visual Studio 2005/2008.
Make sure Visual Studio is closed when installing and complete it.
Open your Visual Studio Project. Right click the solution and select "Add solution to Subversion".
Next type in the Project Name as you wish. Repository URL should be svn://<computername/ipaddress>/myrepos. Click OK.
Now the type in the username+password, and comment. Now a place for your solution is created in the repository. But files are not added.
To add the files, right click the solution and click Commit Solution Changes.
Follow the steps enter a log message again to add all the files.

Now all the files are in repository. To open this project for editing (technical term - check out) in another system - Open Visual Studio 2008, then select File>Open>Subversion Project. Click on the small blue icon Add Repository URL and enter the url again svn://<computername/ipaddress>/myrepos

You can select the solution file inside the directory and check out a working copy to a local folder. Next work on project as you wish, when others have made changes receive them by right clicking solution > Update to Latest Version.

Commit your changes as you progress. Happy team developing !!.

Please leave a comment if you find this useful, thats a big reward I get for spending my time and effort :)

P.S. You have more terms and concept to cover in the SVN book. I will try to write more articles.

Where did the Video control go in Flash CS3

2008-09-24T13:39:00.001+05:30

I have been searching for it. I don't want the FLVPlayback in the components panel. Oh ! god, googling doesn't help. I remember dragging and droppin it from somewhere else long back.. where was it..

Finally go it. How stupid.. I just have to right click in the Library and create a new Video and then drag it to the stage !

Now need to get the demo ready for the Red5 video capture project.. will post more about it soon !

Contact Me

2008-01-02T01:57:00.000+05:30

Hey I am the person who runs this blog. Call me Quakeboy !

Quick Start STL in Visual Studio

2008-01-01T15:32:00.001+05:30

I have searched for a long time to get some practical beginners tutorial for STL in Windows with Visual Studio, but I could not get any useful links which help you kick start

So here something which will really make you kick fly !!!

Declare a map, list etc

#include <list>
#include <map>
using namespace std;

Please dont forget the namespace thing, coz that will never give you proper error message and won't compile !

list<(type)> listname;

e.g.
list<MyClass*> myclassptrslist;

To iterate through the members

list<MyClass*>::iterator myiterator;
for(myiterator= myclassptrslist.begin(); myiterator!=myclassptrslist.end(); myiterator++){

}

You should understand that myiterator is of type MyClass* actually. So to access the original object you can use
(*myiterator).myclassmethod() (or) myiterator->myclassmethod();

Some more pointers

myclassptrslist.front() will give first element
myclassptrslist.back() will give last element and so on

to be continued ..... be right back (brb)

MFC80UD.dll not found ? - A quick solution

2007-11-01T09:49:00.001+05:30

I got this stupid error at random times, I am sure it was a bug, because it came up at random times initially while building and running my MFC Application under VS 2005 Visual Studio .NET

I googled and found some explanations telling that its due to invalid or erraneous manifest file. Well, I really didn't bother to read more and explain about it. I just wanted to make it work for debugging.

I went to WinSxS folder in c:\Windows and found this under the Debug version of MFC dlls. So the application compiled and ran under release mode also.

Solution: Open the project properties > Configuration Properties > General > Project Defaults > Use of MFC = Use MFC in a Static Library

Note : Now you application will run faster (in terms of load times) but will no more be KB's in size, because all the MFC Dll's codes are embedded with it. my >50KB application became 2.9 MB

Overall, there will be ~2.8 MB increase in exe size.

Starting SDL !!!

2007-08-29T17:10:00.000+05:30

Its been quite sometime since I started to study SDL with my friend. I am going to summarise somethings about SDL as a beginner.

First let me tell the steps for configuring SDL with DevC++ IDE 4.9.9.2. After installing DevC++ IDE, download the SDL Library

Download the SDL development library for Mingw32
http://libsdl.org/release/SDL-devel-1.2.12-mingw32.tar.gz
(might be changed in the future)
SDL Library official site is www.libsdl.org

Place it in your C:\ Drive root.

You can also place it whereever you want to store, but here I am going to make further explanation assuming that we are extracting the downloaded file in C:\. If you are extracting at a different location make changes in the path as necessary in the below example

Extract the contents.
It will create a folder called SDL-1.2.12.
Rename the folder to 'SDL' for simplicity sake.
Now extracted contents are in C:\SDL
To verify check if there are folders by name
bin, build-scripts, lib, include etc.
under the path C:\SDL
Now open the DevC++ IDE
Go to Tools Compiler Options.

Under Compiler Tab
Check the 'Add these commands to the linker command line' checkbox.
Then enter
'-lmingw32 -lSDLmain -lSDL'
in that textbox without the single quotes.

Next go to Directories Tab.

Under Libraries add the path C:\SDL\lib
Under C/C++ includes add the path 'C:\SDL\include\SDL'

Now you are all set to begin SDL !!

How to begin game programming ??

2007-08-04T16:59:00.000+05:30

I remember I asked my father how games were made when I was in UKG after I got my ATARI 2600, and has been always trying to find out since then.

(Throughout my school life, I never owned a computer at home, only thing they taught in school was LOGO and very very limited amount of Basic. When I was in 9th or 10th standard I met this awesome thing called "Flash" which was Flash 5 then, and Flash CS4 now !)

This is something that I took years to find out. The first thing I started was to program with Flash. Soon I realized Flash's capabilities.I agree that you can make decent games with flash too. But still how do these people program real AAA titles like Half Life, Quake 3 etc ??

Then I came to world of turbo C, started ambitiously later found its slow and screen was flickering, then came Java with double buffering, I got a tutorial about tile based games and realized this is how games like Super Mario was made. Used it in Java, got some smooth scrolling, but could not make other things happen, too many bugs and I was not clear about many other tile concepts much.

That time my classmate, a super cool brainy guy started out with the same set of tutorials, and made a mixture of Contra, Mario (I had the sprites there !!) But finally the game after all those efforts was too slow. Then I learn t that people make use of something called hardware acceleration to create those super fast high graphics games. I learn t that
there are APIs to use the H/W acceleration and create games.

You can click links to learn more the terms, they all link to Wikipedia pages. Then I started to learn DirectX, the first step in becoming a real game programmer. Fortunately even though I am a diploma holder, I got job in a company to program CAD Applications, I learnt DirectX.NET there and realised it was hard to use it for CAD applications there, then came the world of OpenGL. I liked it mainly because of three things

Its a open industry based standard as opposed to DirectX which was Microsoft's.
John Carmack, my role model uses that to create Quake 3, Doom 3 etc.
Its really simple and powerful, and better than DirectX in many ways.
OpenGL code is very very portable.

Then came SDL, when I started to program games in ubuntu linux. I made a small tetris clone, playable but no scores etc. Remember I told I am lazy.

Now I like SDL, I am planning to learn and make some platformers using it, make is open source and make binaries for linux as well as windows. Lets see whether I am able to do it ???

How to Install Apache Server & MySQL with PHP on Windows XP

2007-02-22T08:43:00.000+05:30

Each time i tried to installed Apache and MySQL on Windows XP with php something or the other doesn't work. So finally I have summarised some working tips.

Download the latest Apache, PHP and MySQL installer for Windows.
First install Apache. Don't forget to enter some email address in the Admin's Mail Address. and its always good to set the port to 80 by default to avoid typing colon and port number each time you access the website.
Then install the MySQL using installer. Select option for Developers Machine, enable root access from remote machines. and then install finally. Sometimes the service doesn't start at the first time. I don't remember why it happens. But finally it starts after two or three tries.
Now extract PHP to any directory you want to. Just make sure the path doesn't have spaces in between. For simplicity sake lets assume its in "e:\php"
Now in the php directory find a file called php.ini-recommended or php.ini-dist. Just make a copy of any one and rename it as php.ini
Now go to the place where you installed Apache. I recommend "c:\apache2" instead of "C:\Program Files\blah blah.....\".
Then find a file called httpd.conf in conf directory of your apache. Open it with notepad
Add the following lines
# For PHP 5 do something like this:
LoadModule php5_module "e:/php/php5apache2_2.dll"
AddType application/x-httpd-php .php
# configure the path to php.ini
PHPIniDir "e:/php"
Now just start the apache server and see if it starts without errors. If done then there is no error installing PHP to Apache.
The second task is to install MySql Extensions to PHP. Here we need two files libmysql.dll (present in the php directory) & php_mysql.dll (present in the ext directory under the php.
Open php.ini and scroll down to find Windows Extensions or just extensions. In php.ini every line beginning with a ';' semi colon is a comment. Uncomment the line extension=php_mysql.dll then scroll up to find and change extension_dir = "e:\php\ext\".
Now you have configured the php.ini still you need to make php find libmysql.dll
You can also find a libmysql.dll in the bin folder of your mysql directory. Use that incase the libmysql which came along with your php didn't work.
Finally copy that libmysql.dll to the system32 subfolder of your windows xp folder. It had worked once without doing this. But I don't know why. Moreover one more option will be to add the bin folder of MySQL to path. It will be done by default while installing and i had disabled it.

Adding a printer in VB.NET

2007-02-16T08:19:00.000+05:30

I took some time for me before I found out that API AddPrinter can be used to add printer whose driver is already installed. All I wanted to do was to add a Printer if its not present and the driver was built-in Windows XP itself.

First I finished adding the printer in VC++ using MFC Application then I wanted to add it using VB.NET either using Declare Function or DllImport by wrapping it in a dll. Then i had to google for a long time when i last arrived at a code that really worked !!!

Module Printer_Module
'declare the API in VB
Private Declare Function AddPrinter Lib "winspool.drv" Alias "AddPrinterW" _
(ByVal pName As IntPtr, ByVal Level As Int32, _
<MarshalAs(UnmanagedType.LPStruct)> ByVal pPrinter As PRINTER_INFO_2) As Int32

'http://www.vbforums.com/archive/index.php/t-263437.html
'still dont know why PRINTER_INFO_2 is Class instead of a Structure

<StructLayout(LayoutKind.Sequential)> _
Public Class PRINTER_INFO_2
<MarshalAs(UnmanagedType.LPTStr)> Public pServerName As String
<MarshalAs(UnmanagedType.LPTStr)> Public pPrinterName As String
<MarshalAs(UnmanagedType.LPTStr)> Public pShareName As String
<MarshalAs(UnmanagedType.LPTStr)> Public pPortName As String
<MarshalAs(UnmanagedType.LPTStr)> Public pDriverName As String
<MarshalAs(UnmanagedType.LPTStr)> Public pComment As String
Public pDevMode As IntPtr
<MarshalAs(UnmanagedType.LPTStr)> Public pLocation As String
<MarshalAs(UnmanagedType.LPTStr)> Public pSepFile As String
<MarshalAs(UnmanagedType.LPTStr)> Public pPrintProcessor As String
<MarshalAs(UnmanagedType.LPTStr)> Public pDatatype As String
<MarshalAs(UnmanagedType.LPTStr)> Public pParameters As String
Public pSecurityDescriptor As IntPtr
Public Attributes As Integer
Public Priority As Integer
Public DefaultPriority As Integer
Public StartTime As Integer
Public UntilTime As Integer
End Class

Function AddMyPrinter() As Boolean

Dim PI2 As New PRINTER_INFO_2()
'only printer name, driver name, port name and print processor need to be filled
With PI2
.pServerName = String.Empty
.pPrinterName = "Apple Color LW 12/660 PS"
.pShareName = String.Empty
.pPortName = "FILE:"
.pDriverName = "Apple Color LW 12/660 PS"
.pComment = String.Empty
.pLocation = String.Empty
.pDevMode = 0
.pSepFile = String.Empty
.pPrintProcessor = "WinPrint"
.pDatatype = String.Empty
.pParameters = String.Empty
.pSecurityDescriptor = 0
.Attributes = 0
.Priority = 0
.DefaultPriority = 0
.StartTime = 0
.UntilTime = 0
End With

Dim x As Integer = AddPrinter(IntPtr.Zero, 2, PI2)
If x = 0 Then
MsgBox("Failed !!!")
MsgBox(Err.LastDllError)
End
End If
Return True
End Function

End Module

Cannot view chm help files !!!

2007-02-16T08:08:00.000+05:30

Suddenly a week ago, all the .chm files (Windows help files) start showing up "This page cannot be displayed error". But all the side links seemed to show up but on clicking them nothing turned up on the main section. First I thought its some kinda virus !! Then i just googled it and obviously many other people have had the same problem. It presented me with a list of scenarios something like an article from microsoft site showed that installation of a update for windows will block these chm files from accessing some places. I just scrolled below for the solution and there was some modification in the registry and i did it and naturally it didn't (coz its microsoft man!!!).

Then i searched and landed up on a blog of some guy who is CEO of some company. The guy had a cool pic of himself and solution was so damn simple. CHM files don't work if you have a '#' hash in the path. for eg. chm file path here was d:\#user\blah.. blah..\file.chm

Then i just copied it to c:\folder\file.chm. folder was without a hash and so it worked.

courtesy : http://www.nik.com.au/archives/2005/04/06/chm-help-files-error-the-page-cannot-be-displayed/
Chm help files error !!!!

DEPLOYING VSTO SOLUTIONS

2006-12-13T12:54:00.000+05:30

This is for all people who know whats VSTO (Visual Studio Tools for office) and have developed apps using it and have not been able to deploy to end user systems. This is the most easy solution I have found to deploy document level solutions i.e. for a single
excel or word file after a long time struggling.

Though there are many ways to deploy VSTO solution for a document level and most handy one i have found so far has been - signing the assembly with a strong name then add entry in the .NET 2.0 security configuration. This is most handy because if you use URL then you cant shift the location of the folder which i needed in my case. Here are the steps.

(NOTE: In the end users computer - You need to install feature .NET forms programmability in Office 2003 by using add/remove features in office 2003 installer and then install VSTOR.exe which can be found in the VSTO CD)
(NOTE: In the developers computer you need the Visual Studio SDK if you need the Microsoft .NET Framework 2.0 Configuration under the admin tools of the control panel)

1) First open your VSTO solution project and open project properties and then go to signing. Then give sign the assembly and key file name choose new. Enter a file name in the dialog box and set password if you need it. then save the project and rebuild it again.
2) Then go to Control Panel>Administrative Tools>Microsoft .NET Framework 2.0 Configuration
3) Then go to Console Root>Microsoft .NET Framework 2.0 Configuration>My Computer>Runtime security policy>Machine>Code Groups>All code>My_Computer_Zone
4) Right click and create new and enter a name click next and then select condition type to Strong Name.
5) Click import below and select the required assembly and check the name and version boxes and give next, provide Fulltrust and then finish.

Now the assembly is ready to run on developers computer anywhere. But you cant follow all the above steps in the endusers computer because you just can't expect every end user to have VS SDK installed. So the Microsoft .NET Framework 2.0 Configuration console doesnt appear.

I had read in many places about caspol.exe tool and tried it but may be i was doing something wrong, and i just couldn't get it working. so the easiest way is this.
In the developers computer open Control Panel>Administrative Tools>Microsoft .NET Framework 2.0 Configuration and thenright click runtime security policy and select create Deployment package, then set the security level to machine and select a place
to save the MSI file and finish it. Now this msi file can be used to automatically follow the steps 2-5 in the endusers computer.

So dont forget to restart the system after running the MSI file in the end users computer. If any doubt, comment and it will reply or mail me........