I went down a few algorithmic roads recently, digging into path finding and – for some obscure reasons – bit manipulations. Or byte. Whatever.

Along this way some utility methods (or functions) were born, and I thought: May be some day some of them may be in use to any of you ;)

For my dear non-geeky readers: A bit is the smallest part in software. It’s either this or that, either 0 or 1, either false or true. With a group of 2 bits you already have 4 states: 00, 01, 10 and 11. With 8 it’s 256 and so on (2^n).

As it would be too boring to just type 0 or 1, and because we have more than 2 fingers, man invented numbers to accumulate these bits: so 9 stands for 1001, and because 9 is shorter than 1001, we prefer 9. Some even write AB for 10101011, but that’s where we come back to geeky world.

So after this highly informative introduction, let’s get to some code. First, let’s count bits:

static public function countBits( value : uint ) : uint {
	var count:uint = 0;
	while (value) {
		if ( value & 1 ) {
			count++;
		}
		value >>>= 1;
	}
	return count;
}

Example:
countBits( 0xAB ) -> 5

Now sometimes you might wanna know: Does this data contain no more than 1 bit? We could just ask countBits( value ) == 1. But that’s not as speedy as it should be, right? So here we go:

static public function is_1_bit( value : uint ) : Boolean {
	var count:uint = 0;
	while (value) {
		if ( value & 1 ) {
			if (count == 1) return false;
			count++;
		}
		value >>>= 1;
	}
	return count == 1;
}

Examples:
is_1_bit( 0xAB ) -> false
is_1_bit( 0×400 ) -> true

uint are by the way 32bit data, so a maximum of 32 of these bits we’re talking about can be turned on or off. That’s a lot of data. 4’294’967’296 combinations (though not that high compared to the numbers we read every day in the newspapers recently). Anyway, sometimes we might wanna access and set only a group of bits (usually 4 or 8) within this quite large row of bits:

static public function getBitGroup( value : uint , group : uint , len : uint = 4 ) : uint {
	return ( value >> (group*len) ) % (1 << len);
}
 
static public function setBitGroup( value : uint , groupValue :uint , group : uint , len : uint = 4 ) : uint {
	var pos:uint = group * len;
	var mask:uint = n_bits(pos);
	var right_bits:uint = value & mask;
	value >>>= pos + len;
	value <<= len;
	value |= groupValue;
	value <<= pos;
	value |= right_bits;
	return value;
}

Don’t they look just groovy?! Yeah baby!

Anyway, that’s all for now. Stay tuned for some crazy path finding. If I find time (sometimes I wonder how all those bloggers find their time to write so much..) Not to mention Twitter. Boohaa.

Related posts:

1. Simplicity follows Performance – parsing Strings in Actionscript 3
2. Metatunnel 1k Demo: AS vs. JS

There are lots of articles written about the garbage collector and I’m not going into it any deeper. Let’s just summarize that no developer likes that kind of behaviour — it’s totally unpredictable. System.gc() would help a little, but it’s only available to debug players.

You may say: what do you care about memory handling! And I’d answer: not that much actually! :-) But what I really care about is false behaviour that can result.

Within Flash we have two ways to keep weak references to objects: Dictionary and weak listeners (weak method closures). We use weak references so that objects will be collected by the garbage collector. Now when it comes to Dictionaries, they behave as I’d expect. A “dead” object won’t be listed in a for each loop. But events events events…. they’ll be dispatched to each and every “dead” object residing in memory!! Which is such a pain in the ass really!

After a lot of testing I can give only the advice you’ve probably heard many times before:

Always remove listeners! Even the weak ones!

Otherwise you have to potentially deal with unexpected behaviour. I may gonna create some utility class for that that deals with this problem.

Here my testing code:

package
{
	import flash.display.Sprite;
	import flash.events.Event;
	import flash.events.MouseEvent;
	import flash.events.TimerEvent;
	import flash.system.System;
	import flash.text.TextField;
	import flash.text.TextFieldAutoSize;
	import flash.text.TextFormat;
	import flash.utils.Dictionary;
	import flash.utils.Timer;
 
	import net.hires.debug.Stats;
 
	public class MemoryLeakTest extends Sprite
	{
		public static var counter : uint;
		public static var dict:Dictionary = new Dictionary( true );
 
		private var t:TextField;
		private var makeGC:Boolean;
		private var switcher:uint;
 
		public function MemoryLeakTest()
		{
			super();
 
			t = new TextField();
			t.defaultTextFormat = new TextFormat( null, 18 , 0x999999 , true );
			t.autoSize = TextFieldAutoSize.LEFT;
			addChild(t);
 
			var timer : Timer = new Timer( 100 );
			timer.addEventListener(TimerEvent.TIMER,handleTimer);
			timer.start();
 
			stage.addEventListener(MouseEvent.CLICK,handleClick);
 
 
			stage.addChild( new Stats() );
		}
 
		public function handleTimer( event : TimerEvent ) : void {
			counter = 0;
			if ( makeGC ) System.gc();
			dispatchEvent( new Event( Event.CHANGE ) );
 
			var i:uint;
			for each ( var obj:* in dict ) {
				i++;
			}
 
			t.text = ( "Still " + counter + " dead objects living... (GC " + (makeGC?'on':'off') + ')' );
			t.appendText( '\nIn dictionary: ' + i );
			t.appendText( '\nClick to switch manual Garbage Collection aka System.gc()' );
 
			if ( switcher++ % 2 ) {
				new TestSprite( this );
				t.appendText( '\nAdded new test object (should live for 10 msec)' );
			}
		}
		public function handleClick( event : Event ) : void {
			makeGC = !makeGC;
		}
	}	
}
import flash.display.Sprite;
import flash.display.InteractiveObject;
import flash.utils.Timer;
import flash.events.TimerEvent;
import flash.events.Event;
import flash.system.System;
 
 
 
class TestSprite extends Sprite {
 
	// the TestSprite instance should not live longer than 10 msec
 
 
	public function TestSprite( parent : Sprite ) : void {
		// add to weak dictionary
		//MemoryLeakTest.dict[ this ] = true;
 
		// temporarily add to parent
		parent.addChild( this );
		var timer : Timer = new Timer( 10 );
		timer.addEventListener(TimerEvent.TIMER,handleTimer);
		timer.start();
 
		// listen (weak!!) to parent event - this is just for control purposes
		parent.addEventListener( Event.CHANGE , handleChange , false , 0 , true );
 
		// let's have those sprites eat some memory while alive
		this.cacheAsBitmap = true;
		graphics.beginFill(0xff0000,0.1);
		graphics.drawRect(0,0,500,500);
		graphics.endFill();
	}
	public function handleTimer( event : TimerEvent ) : void {
		( event.target as Timer ).removeEventListener(TimerEvent.TIMER,handleTimer);
		parent.removeChild(this);
	}
	public function handleChange( event : Event ) : void {
		MemoryLeakTest.counter++;
	}
}

No related posts.

Picture by Marc Liyanage

It’s been surprisingly fun (I give credits to the beer sponsored by Nemos). People even managed to pretend they’d be interested in what I was prosing, so credit to them too!

I’d also like to mention that there’ll be the swiss flash event soon: Flash at the Lake will not only pamper you with appearances of great national and international Flash enthustiacs, it will also give anybody attending the pleasure to enjoy one of Zurich’s greatest locations with people who don’t think of you as a storm lightning adorer when you sit in the sun twittering Flash into the clear lake. And all that to a fantastic price. Check it out at fatl.ch

So here we go with the presentation (Quicktime so you can enjoy the marvelous effects):

No related posts.

Yeah well, I was more than optimistic to show those JS guys how fast Flash can be with the help of some brand new Adobe magic – but Pixel Bender was, unfortunately, quite disappointing:

1. I spent hours to successfully migrate the code to Pixel Bender Toolkit (wasn’t that difficult actually, but a wrong character there and your math is all screwed up). So here the first letdown: Pixel Bender Toolkit is not a very nice coding environment. It’s a rough baby. But I’d do everything for speed :-)..
2. Once I had it running (the above picture is captured from Pixel Bender) I tried to export it for Flash Player. But hoohoo! no support for loops in Flash Player! Yeah, well, ..(and also no custom functions by the way)
3. But I didn’t give up. I just calculated how many max loops where needed (about 100) and replicated the while loop with many if-conditions (I love to do computer’s dummy job ;). Then finally: Export to Flash Player! But in Flash Player having the shader as Filter for a BitmapData just really fucked up the rendering. I even went down to 1fps, but still no luck. What a bummer! (sometimes I saw a few pixels blue, but very unusable: I’ve written to the Pixel Bender Forum - invited by @Pixelbender – but that led to nothing so far).
4. Today, finally, I got it working by taking a more rough approach (not BitmapFilter, but ShaderJob). But hell is it slow!!! And it eats all my 8 cpu cores!! So I wondered how that could be: Fast as hell in Pixel Bender Toolkit (uses may be 2% of my cpu at full fps!), but slow as hell in Flash Player?! The answer is obvious: Adobe decided not to talk to the GCU (Graphics Card) for the calculations, probably to keep the Flash Player be as platform independent as possible – and as small as possible. But then I wonder.. WTF do you give us this toy if we can’t use it?!! It’s like Apple would say: «Great news: we have Core Graphics on the iPhone – without hardware acceleration..» Adobe, to me, this doesn’t make too much sense. (But I’m so very much pleased about the new Text Engine – *that* was a good job)

Here the result: (Click to start, it’ll eat your cpu!)

And for the geeks, here we go with the source codes:

<languageVersion : 1.0;>
 
kernel MetaTunnel
<   namespace : "com.betabong";
    vendor : "Betabong";
    version : 1;
    description : "MetaTunnel Port";
>
 
{
    parameter int size
    <
        minValue: 16;
        maxValue: 512;
        defaultValue: 128;
    >;
 
 
    parameter float time
    <
        minValue: float( 0.0 );
        maxValue: float( 15.0 );
        defaultValue: float( 0.0 );
    >;
 
 
 
 
    input image4 src;
    output pixel3 dst;
 
    void
    evaluatePixel()
    {
        float cos1 = cos( time );
        float cos0_5 = cos(time * 0.5);
        float cos0_7 = cos(time * 0.7);
        float sin1 = sin( time );
        float sin0_2 = sin(time * 0.2);
        float sin_0_5 = sin(time * 0.5);
 
        float dim = float( size );
        float2 point = outCoord() / dim;
        if ( point.x > 1.0 || point.y > 1.0 ) {
            dst.rgb = float3( 1.0 , 1.0 , 1.0 );
        } else {
 
        float2 vp = float2( point.x * 2.0 - 1.0 , -point.y * 2.0 + 1.0 );
        float s = float( 0.4 );
        float3 op = float3( vp.x , vp.y * 1.25 , 0.0 );
        float3 dp = float3( (vp.x + cos1 * 0.3) / 64.0 , vp.y/64.0 , 1.0/64.0 );
        float f = float( 1.0 );
 
        float tt = float( 0.0 );
        float g = float( 1.0 );
        float3 p = float3(1.0,1.0,1.0);
        while ( (g > s) && (tt<375.0) ) {
            p = op + ( dp * tt );
            f = 1.0;
            f *= abs( distance( float3( cos1+sin0_2 , 0.3 , 2.0+cos0_5*0.5 ) , p ) );
            f *= abs( distance( float3( -cos0_7 , 0.3 , 2.0+sin_0_5 ) , p ) );
            f *= abs( distance( float3( -sin0_2*0.5 , sin1 , 2.0 ) , p ) );
            f *= cos(p.y)*cos(p.x) - 0.1 - cos( p.z*7.0 + time*7.0 ) *cos(p.x*3.0)*cos(p.y*4.0)*0.1;
            g = f;
            tt += g * 4.0;
        }
        float color = 0.0;
        float3 dtt = op + ( dp * tt);
 
        p = float3( dtt.x , dtt.y , dtt.z );
        f = 1.0;
        f *= distance( float3( cos1+sin0_2 , 0.3 , 2.0+cos0_5*0.5 ) , p );
        f *= distance( float3( -cos0_7 , 0.3 , 2.0+sin_0_5 ) , p );
            f *= distance( float3( -sin0_2*0.5 , sin1 , 2.0 ) , p );
        f *= cos(p.y)*cos(p.x)-0.1-cos(p.z*7.0+time*7.0)*cos(p.x*3.0)*cos(p.y*4.0)*0.1;
        float objd = f;
 
        float3 np = float3( 0.0,0.0,0.0 );
 
        p = float3( dtt.x + 0.01 , dtt.y , dtt.z );
        f = 1.0;
        f *= distance( float3( cos1+sin0_2 , 0.3 , 2.0+cos0_5*0.5 ) , p );
        f *= distance( float3( -cos0_7 , 0.3 , 2.0+sin_0_5 ) , p );
            f *= distance( float3( -sin0_2*0.5 , sin1 , 2.0 ) , p );
        f *= cos(p.y)*cos(p.x)-0.1-cos(p.z*7.0+time*7.0)*cos(p.x*3.0)*cos(p.y*4.0)*0.1;
        np.x = objd - f;
 
        p = float3( dtt.x , dtt.y + 0.01 , dtt.z );
        f = 1.0;
        f *= distance( float3( cos1+sin0_2 , 0.3 , 2.0+cos0_5*0.5 ) , p );
        f *= distance( float3( -cos0_7 , 0.3 , 2.0+sin_0_5 ) , p );
            f *= distance( float3( -sin0_2*0.5 , sin1 , 2.0 ) , p );
        f *= cos(p.y)*cos(p.x)-0.1-cos(p.z*7.0+time*7.0)*cos(p.x*3.0)*cos(p.y*4.0)*0.1;
        np.y = objd - f;
 
        p = float3( dtt.x , dtt.y , dtt.z + 0.01 );
        f = 1.0;
        f *= distance( float3( cos1+sin0_2 , 0.3 , 2.0+cos0_5*0.5 ) , p );
        f *= distance( float3( -cos0_7 , 0.3 , 2.0+sin_0_5 ) , p );
            f *= distance( float3( -sin0_2*0.5 , sin1 , 2.0 ) , p );
        f *= cos(p.y)*cos(p.x)-0.1-cos(p.z*7.0+time*7.0)*cos(p.x*3.0)*cos(p.y*4.0)*0.1;
        np.z = objd - f;
 
        float d = length( np );
        np.y /= d;
        np.z /= d;
 
        color = max( -0.5 * np.z , 0.0 ) + max( -0.5 *np.y + 0.5 * np.z , 0.0 ) * 0.5;
        float3 rgb = float3(color  + 0.1 * tt * 0.025 , color  + 0.2 * tt * 0.025  , color + 0.5 * tt * 0.025   );
        rgb.x = max( 0.0 , min( 1.0 , rgb.x ) );
        rgb.y = max( 0.0 , min( 1.0 , rgb.y ) );
        rgb.z = max( 0.0 , min( 1.0 , rgb.z ) );
 
        dst.rgb = rgb;
 
 
      }
 
    }
 
}

And here the Actionscript part:

package
 {
    import flash.display.*;
    import flash.events.*;
    import flash.filters.*;
    import flash.net.*;
    import flash.utils.ByteArray;
 
    import net.hires.debug.Stats;
    // SWF Metadata 
    [SWF(width = "520", height = "128", backgroundColor = "#000000", framerate = "1")]
    public class MetaTunnelBender extends Sprite {
 
        [Embed("test/MetaTunnel-stupid.pbj", mimeType = "application/octet-stream")]
        private var TestFilter: Class;
        private var playing : Boolean;
 
        private var im: Bitmap;
        private var job:ShaderJob;
        private var shader: Shader;
        private var shaderfilter:ByteArray;
        private var time:Number = 0.0;
        private const SIZE:Number = 64;
        private const WHITE:BitmapData = new BitmapData( SIZE , SIZE , false , 0xffffff );
 
        public function MetaTunnelBender() : void {
        	stage.frameRate = 20;
            im = new Bitmap( WHITE.clone() );
            im.scaleX = im.scaleY = 2;
            addChild(im);
 
            // stat info
            var s:DisplayObject = new Stats();
            addChild( s );
            s.x = 128;
 
            //Pass the embedded filter to the Shader as a ByteArray 
            shaderfilter = new TestFilter() as ByteArray;
            shader = new Shader( shaderfilter );
            shader.data.size.value = [ SIZE ];
 
			//start shaderjob (async start won't work - bug?)            
            job = new ShaderJob( shader , im.bitmapData , im.width , im.height );
			job.start( true );
 
			// let's start/stop on click
			stage.addEventListener( MouseEvent.CLICK , click );
 
            // render first frame
			enterframe();
        }
 
        private function enterframe( e:Event = null ) : void {
        	time += 0.05;
			shader.data.time.value = [ time ];
			trace( job.progress );
            job = new ShaderJob( shader , im.bitmapData , im.width , im.height );
			job.start( true );
       }
 
        private function click( e:* = null ):void {
        	if( !playing ) {
	           addEventListener( Event.ENTER_FRAME , enterframe );
	        } else {
	           removeEventListener( Event.ENTER_FRAME , enterframe );
	        }
	        playing = !playing;
        }
 
    }
 
}

Related posts:

1. Metatunnel 1k Demo: AS vs. JS
2. GridFitType (& animation)

Paulo Falcão ported this to Javascript using canvas.

To make the set complete I ported Paulos JS version to Actionscript, just quick’n'dirty.

Click on it to start the animation:

Here’s the code:

package test
{
	import flash.display.Bitmap;
	import flash.display.BitmapData;
	import flash.display.Sprite;
	import flash.events.Event;
	import flash.events.MouseEvent;
 
	public class JSCanvasPort extends Sprite
	{
		public function JSCanvasPort()
		{
			super();
			addChild( canvas = new Bitmap() );
			canvas.smoothing = false;
			canvas.scaleX = canvas.scaleY = 2;
			draw();
 
			addEventListener( MouseEvent.CLICK , handleClick );
		}
 
		private var playing:Boolean = false;
 
		public function handleClick( e:Event = null ) : void {
			if ( playing )
				removeEventListener( Event.ENTER_FRAME , enterframe );
			else
				addEventListener( Event.ENTER_FRAME , enterframe );
			playing = !playing;
		}
		public function enterframe( e:Event=null ) : void {
			draw();
		}
 
		public var time:Number=0;
		public var maxr:Number=64.0;
		public var canvas:Bitmap;
 
		public function distance(ax:Number,ay:Number,az:Number,bx:Number,by:Number,bz:Number):Number{
			var dx:Number=bx-ax , dy:Number=by-ay , dz:Number=bz-az;
		    return Math.sqrt( dx*dx + dy*dy + dz*dz );
		}
		public function obj(x:Number,y:Number,z:Number,t:Number):Number{
		        var f:Number=1.0;
		        f*=distance(x,y,z,Math.cos(t)+Math.sin(t*0.2),0.3,2.0+Math.cos(t*0.5)*0.5);
		        f*=distance(x,y,z,-Math.cos(t*0.7),0.3,2.0+Math.sin(t*0.5));
		        f*=distance(x,y,z,-Math.sin(t*0.2)*0.5,Math.sin(t),2.0);
		        f*=Math.cos(y)*Math.cos(x)-0.1-Math.cos(z*7.0+t*7.0)*Math.cos(x*3.0)*Math.cos(y*4.0)*0.1;
		        return f;
		}
 
		public function eval(x:Number,y:Number,t:Number):uint{
		   var vx:Number=x*2.0-1.0; var vy:Number=-y*2.0+1.0;
		   var s:Number=0.4;
		   var ox:Number=vx;var oy:Number=vy*1.25;var oz:Number=0.0;
		   var dx:Number=(vx+Math.cos(t)*0.3)/64.0;var dy:Number=vy/64.0;var dz:Number=1.0/64.0;
		   var tt:Number=0.0;
		   var g:Number=1.0;
		   while((g>s)&&(tt<375)){
		         g=obj(ox+dx*tt,oy+dy*tt,oz+dz*tt,t);
		         tt+=g*4;
		   };
		   var color:Number=0.0;
		   var dxtt:Number=ox+dx*tt;var dytt:Number=oy+dy*tt;var dztt:Number=oz+dz*tt;
		   var objd:Number=obj(dxtt,dytt,dztt,t);
		   var nx:Number=objd-obj(dxtt+0.01,dytt,dztt,t);
		   var ny:Number=objd-obj(dxtt,dytt+0.01,dztt,t);
		   var nz:Number=objd-obj(dxtt,dytt,dztt+0.01,t);
		   var d:Number=Math.sqrt(nx*nx+ny*ny+nz*nz);ny=ny/d;nz=nz/d;
		   color+=Math.max(-0.5*nz,0.0)+Math.max(-0.5*ny+0.5*nz,0.0)*0.5;
		   var r:Number=(color+0.1*tt*0.025);
		   var g:Number=(color+0.2*tt*0.025);
		   var b:Number=(color+0.5*tt*0.025);
		   return getRGB(r,g,b);
		}
 
		public function getRGB(r:Number,g:Number,b:Number):uint{
		        if (r<0) r=0; else if (r>1) r=1;
		        if (g<0) g=0; else if (g>1) g=1;
		        if (b<0) b=0; else if (b>1) b=1;
			return ((r*255)&255) << 16 | ((g*255)&255) << 8 | ((b*255)&255);
		}
 
		public function draw():void{
			if ( canvas.bitmapData == null ) {
				canvas.bitmapData = new BitmapData( maxr , maxr , false , 0xffffff );
			}
			var t:BitmapData = canvas.bitmapData;
			time+=0.1;
			t.lock();
		        for(var x:Number=0;x<maxr;x++){
		                var px:Number=x/maxr;
		                for(var y:Number=0;y<maxr;y++){
		                        var py:Number=y/maxr;
		                        t.setPixel( x , y , eval(px,py,time) );
		                }
		        }
		   t.unlock();
		};
 
	}
}

Update: A quickly optimized version (nothing advanced really).

(You can switch quality in this version with key up or down).

Related posts:

1. MetaTunnel with Pixel Bender
2. Weak Method Closure
3. 3D Point Cloud with Javascript and Canvas

Well not really much to say here:

• NONE Good for animation and people who prefer Font appearance over readability
• SUBPIXEL Good compromise between readability (small sizes) and appearance
• PIXEL Pure nonsense if you ask me

What I really wonder though is: Why on earth can’t I adjust fractioned text sizes (aka float)??? In the IDE, yes you can. By actionscript, no you can’t. I don’t get it.. anybody knows why that is? And more importantly: is there some fancy workaround? (in the example I’m just scaling the text fields, but that’s not really a cool solution).

No related posts.

It happens to DisplayObjects A inside DisplayObjectContainers B inside DisplayObjectContainer C, when

• A was not initially visible (not inside initial scroll rect of C)
• A is in 3D mode (I just change rotationY for that)
• B is in «cached as Bitmap» (cacheAsBitmap would do, I go with DropShadowFilter in the example)
• C’s scrollrect property is set, so A is shows up (well, it doesn’t – that’s the bug after all ;)

Here the example:

bug-3d-shadow.swf

And here the code:

/* Flash Bug - Flash Player 10.0.22.87
	disappearance when 3d's parent has shadow
	more info: betabong@gmail.com
*/
 
this.stage.scaleMode = StageScaleMode.NO_SCALE;
this.stage.align = StageAlign.TOP_LEFT;
 
 
var scroller:Sprite = new Sprite();
addChild( scroller );
scroller = this;
 
var filter:DropShadowFilter = new DropShadowFilter( 2 , 90 , 0 , 0.5 , 8 , 8 , 1 , 2 );
var r:Sprite,rect:Sprite;
 
for ( var mode:int = 0 ; mode<=1 ; mode++ ) {
	for ( var i:int=0 ; i<3000 ; i+=100 ) {
		r = new Sprite();
		r.addChild( rect = new Sprite() );
		rect.graphics.beginFill( 0xff0000 );
		rect.graphics.drawRect( -40 , -40 , 80 , 80 );
		rect.graphics.endFill();
		rect.addEventListener( MouseEvent.MOUSE_MOVE , rotate );
		if ( !mode ) r.cacheAsBitmap = true; //r.filters = [ filter ];
		r.y = i + 100;
		r.x = mode * 150 + 100;
		addChild( r );
	}
}
 
stage.addEventListener( Event.ENTER_FRAME , move );
 
var scroll:Number = 0;
function move( event:Event ) : void {
	scroll += ( stage.mouseY - 100 ) / 50;
	scroller.scrollRect = new Rectangle( 0 , scroll , 500 , 200 );
}
 
 
function rotate(e:*) {
	e.target.rotationY++;
}

Update: Now that: it took me about 20 min to file a bug in the Adobe Bug System (which is slow as hell anyway). And when I want to submit it, I get a message saying they’d be in maintenance. Well, buggy bug systems piss me off quite a bit.

No related posts.

Though I’ve written some posts about how to code within Flex Builder and compile from there using Flash IDE, I personally don’t like that at all and only use it for some few older AS3 projects. There are better methods, ways to compile from within Flex Builder while still being able to make use of Flash comfort.

There’s a neat way to embed a library without loosing any functionality (like little scripts). I’ve first seen it at Grant Skinner (who’s doing great stuff, one of my favorites in the flash community really). Let’s say, you have a assets.fla and a published assets.swf. Now here we go:

[SWF(width="750", height="500", frameRate="30", backgroundColor="#111111")]
dynamic public class MyApplication extends Sprite
{
 
	[Embed(source="assets/assets.swf", mimeType="application/octet-stream")]
	public var StageAssets:Class;
 
	public function GlobusFlow()
	{
		loadEmbedded( StageAssets );
	}
 
	public function loadEmbedded( cls : Class ) : void {
		trace( "[Preloader] Loading embedded" );
		var loader : Loader = new Loader();
		loader.contentLoaderInfo.addEventListener( Event.COMPLETE , this.handleComplete );
		var context : LoaderContext = new LoaderContext( false , ApplicationDomain.currentDomain );
		loader.loadBytes( new cls() , context );
	}
 
	protected function handleComplete( e : Event ) : void {
		var app : Sprite = Sprite( LoaderInfo( e.target ).content );
		this.loaded( app );
	}
 
	protected function loaded(target:Sprite):void {
		var background : MovieClip = target as MovieClip;
 
		var mainChildren : Array = [];
		for ( var i:int = 0 ; i<background.numChildren ; i++ ) {
			mainChildren.push( background.getChildAt(i) );
		}
		for each ( var child : DisplayObject in mainChildren ) {
			if ( child.name ) {
				this[ child.name ] = child;
			}
			this.addChild( child );
		}
 
		stage.stageWidth = background.width;
		stage.stageHeight = background.height;
	}
}

This does a bit more than just make the library available. In loaded() we’ll loop through all assets that are «on stage» in the loaded/embedded movieclip (assets.swf), and add each of them to our stage. That leads to a state that’s quite similar to as if we’d code within assets.fla itself. Hmm… well, really, that sounds more complicated than it is :-) It mainly gives you the feeling of being able to code within a SWF file. You can visually do your stuff in assets.fla, you can add a background and whatever to its stage, keep a nice library with graphics, animations, sounds and fonts.. and schwupps, here you go!

If you have Fonts embedded in your library, there’s a tiny little more to do:

protected function loaded(target:Sprite):void {
	// ... (all the other stuff, see above)
	registerFont( 'BoldFont' );
}
 
public static function registerFont( fontLinkageId : String ) : void {
	var fontClass : Class = getDefinitionByName( fontLinkageId ) as Class;
	Font.registerFont( fontClass );
}

If you just wanna embed fonts without any registering, you can also go like that:

[Embed(source='assets/fonts.swf' , symbol='BoldFont')]
public var BoldFont : Class;

Of course, instead of embedding everything, you can also load this dynamically (I mean, after all we have the Loader ready, don’t we? :) The advantage of that is that you can give your assets to some dumb Flasher and he can mess around with it without you having to recompile the app.

I know, this is all snippets and frickets, I’m just too lazy right now to build an example Flex project. Still I hope it’s of some use for one or the other.

Related posts:

1. Test Movie from Flex to Flash (easy way)
2. Test Flex/Flash throttled (aka Simulate Download)

The thing with the old Papervision version was.. it got pretty slow when loading many pictures (resulting in many 3d objects). I got briefed with a screen design more or less and expected much less elements to be loaded. Now they got back to me saying they had some problems. Oh yes, they had. Boy was it slow!!

So I did a new version, skipping Papervision for the new Flash Player 10 capabilities. Wow! What an improvement not only in speed, but also in quality (because I wouldn’t have to switch to crispy bitmap handling for acceptable speed – and Flash doesn’t provide that feature anyway) and file size (less than 25%!). Now I just hope that the end client will have the nuts to go with Flash Player 10 :-) (Adobe still doesn’t provide any statistics, but it’s not that wide spread yet as for now).

See here the result (still in development – click the picture). You’ll need Flash Player 10 (I don’t check for the version in this example):

I made use of SimpleZSorter

Update: Well, the client wouldn’t go with 10. So I had to recode the Papervision version to speed it up as much as possible, and I think it got pretty speedy after all. See here for comparison.

No related posts.

Once you got used to developing for Flash in Flex Builder, you hate to do any programming in Flash IDE. Still you sometimes might have to: Flex Builder won’t allow to publish into a FLA file. (There are of course many other reasons, like supporting older AS1/AS2 projects, we don’t go into that here.) Thanks to Eclipse’s ability to be customized, there are ways to make things at least a little easier. I give here a little overview of how I’ve set up my environment, based on several helpful resources I’ve found in the web.

1. Install FlashCommand
2. Install Ant

And for each project:

1. Create/modify Ant file
2. Create Actionscript project
3. Set up project and FLA file

What you need

• Flex Builder
• Flash IDE (CS3 or CS4)

Install FlashCommand

flashcommand is a python script by Mike Chambers that lets you compile FLA files from the command line using Flash IDE. Download it and put it wherever you want.

Install ANT

If you’re used to FDT (or Eclipse in any way) you may well know Ant. Ant is the premier build tool for any developers working in Eclipse, with many many cool possibilites to automate your workflow. Here’s how you install Ant in Flex Builder 3 Standalone:

1. Go to Help > Software Updates > Find and Install
2. Search for new features to install, click next
3. Select «The Eclipse Project Updates», click finish
   Note: If you do not have the option above click «New Remote Site» and enter «The Eclipse Project Updates» as the name and «http://update.eclipse.org/updates/3.3» as the url.
4. In Eclipse project updates look for «Eclipse Java Development Tools …» – it might be in «Eclipse SDK Eclipse 3.3.2» but this depends on what version you have installed and what version is currently available. Select it and click next.
5. Accept licence agreement, click next. Then click «Finish» to start download.
6. Once downloaded, click «Install all»
7. Restart Eclipse (you’ll be asked to)

Create Project

1. Create a new Actionscript Project in Flex Builder
2. Add a new «build» directory in the project root
3. Add your FLA file(s) and assets to the build folder and your source files to the src folder
   (I also strongly recommend adding external libraries to and «externals» directory and use svn:externals property on it, but that’s another story)
4. Add build paths. You have to do this at least twice:
   • In the Flash IDE for the FLA: Publish Settings > Actionscript 3 «Settings…». Don’t forget to add the src folder at least.
     Test your FLA once done. It should compile just fine.
   • In Flex Builder: Project > Properties > ActionScript Build Path. Add your external libraries here. This is used so Flex knows where to lookup your classes, it won’t actually be needed for compiling, because that’s what Flash IDE’s gonna do.
5. Now you’re almost done: you can edit your classes in Flex Builder and switch to Flash IDE to publish and/or test your movie. We’re gonna make this a little more comfortable with Ant and flashcommand.

Add Ant build file

1. Create a new file build.xml in the root folder and open it with Ant Editor (in Flex Builder aka Eclipse)
2. Edit the build.xml according your setup (see below for example)
3. Add build.xml to Ant View (Window > Other Views…)
4. Run an action by double clicking it in Ant View: there you go!! It will compile in Flash IDE and open in your browser for testing!

Sample Ant build.xml

	<?xml version="1.0" encoding="UTF-8"?>
	<!-- ====================================================================== 
	     Nov 28, 2008 1:37:54 AM                                                        
 
	     MyProject    
	     Compile in Flash IDE
 
	     sev                                                                
	     ====================================================================== -->
	<project name="MyProject" default="Compile and Run" basedir="./">
	        <!-- project specifics -->
	        <property name="browser" value="Safari"/>
	        <property name="swf.filename" value="MyProject"/>
	        <property name="html.filename" value="fla/index.html" />
 
	        <!-- Flash IDE Commandline compiler -->
	        <property name="python" value="/usr/bin/python" />
	        <property name="flash.command" value="/path/to/flashcommand" />
 
	        <!-- directories -->
			<property name="user.path" value="/Users/yourname"/>
 
	        <property name="build.dir" value="${basedir}/build" />
 
	        <property name="main.fla" value="${build.dir}/${swf.filename}.fla" />
	        <!-- paths -->
	        <property name="output.path" value="${build.dir}/${swf.filename}.swf" />
 
			<!-- you can also point to the file:///path/to/index.html -->
	        <property name="local.path" value="http://localhost/projects/MyProject/${html.filename}"/>
	        <property name="log.path" value="${user.path}/Library/Preferences/Macromedia/Flash Player/Logs/flashlog.txt"/>
 
	        <!-- Compile python flashcommand -e  -s build/MyProject.fla -o build/MyProject.swf 
 
	        usage: flashcommand -e | -c | -p [-v] [-x] (-s <sourcefile>) ([-o] <exportpath>) ([-t] <timeout>)([-d] <tempdir>) [-j]
 
	        Options and arguments:
 
	        -a : Prints version and about information.
	        -c : Specifies save and compact action. 
	        -d : Specifies temp directory that will be used for temporary files. Optional.
	        -e : Specifies export action.
	        -h : Prints usage information.
	        -j : Specifies that the generated JSFL file should be printed. If this option is specified, Flash will not be executed.
	        -o : Specifies the output file if -e flag is also set. Optional. If not specified, file will be output to same directory as source.
	        -p : Specifies publish action.
	        -s : Specifies source file. Required.
	        -t : Specifies timeout value. Optional.
	        -v : Specifies verbose mode. Optional.
	        -f : Specifies that the Flash authoring version installed is a version other than Flash CS3
	        -x : Specifies whether Flash should be closed after it is done processing. Optional.    
 
	        -->
 
	<target name="Compile and Run">
		<antcall target="Compile in IDE"/>
		<antcall target="Clear Log"/>
		<antcall target="Open in Browser"/>
	</target>
 
        <target name="Compile in IDE">
        		<echo>${flash.command} -e -c -s ${main.fla} -o ${output.path}</echo>
                <exec executable="${python}" failonerror="true" logError="true">
                        <arg line="${flash.command}" />
                        <arg line="-e " />
                        <arg line="-s ${main.fla} "/>
                        <arg line="-o ${output.path}"/>
                </exec>
        </target>
 
 
        <!-- Open in local browser -->
        <target name="Open in Browser">
                <exec executable="open">
                        <arg line="-a ${browser} ${local.path}" />
                </exec>
        </target>
 
        <!-- Clear Debug Log -->
        <target name="Clear Log">
 
        	<concat destfile="${log.path}">                                                                                                            
.: Sev Log File :.
 
</concat>
            <exec executable="open">
                    <arg line="-a console '${log.path}'" />
            </exec>
        </target>
	</project>

Update: Lee Brimelov just released a plugin for Eclipse, that provides this functionality (more or less) with a single click of an icon. It’s an alpha version, and you won’t have all the Ant flexibility (well, you could still install Ant), but you can avoid most of the installation hazzle, and that’s pretty cool. Unfortunately I couldn’t get it to work so far.

Update 2: It bugged me that every time I execute this action Flash would pop up. To work around that I slightly modified the FlashCommand script (it actually checks wether the FLA file is already open, and if so won’t call the open command). Also I had to leave away the -c option (save and compact), something I don’t need anyway. So now when you execute the command from Flex the following happens:

1. The little publish dialog pops up (Fla is exported to SWF)
2. The HTML file is opened in the browser (which comes to front), as well as the Console displaying the trace output

Also I have bound the «Launch last external tool» command in Flex Builder to a key combo. Now things run pretty smooth and as much satisfying as compiling via Flash can be. Btw, I also looked into this Eclipse plugin – as it seems the only thing it does is to tell Flash to «Test Movie» with the frontmost open document. That’s absolutely okay for most people, but I’m glad for the extra functionality I get with this solution (at the cost of more initial effort of course). Download here the modified FlashCommand: flashcommand.zip

Related posts:

1. Test Movie from Flex to Flash (easy way)
2. Test Flex/Flash throttled (aka Simulate Download)
3. Flex Project – SVN Strategy

This has been sneak peaked a while ago (see Video below) – now it’s here. And I’m amazed!! Think of the possiblities: XSLT interpreter (as demoed in the vid), 3D speed, super fast parsing, physics, AI – everything that’s CPU intensive and won’t need any platform specific stuff.

So far this C code still looks pretty fucked up to me (see examples at the labs), but hell, I’m sure that one day we’re gonna see some real fancy stuff thanks to this. I’m looking forward!

Related posts:

1. Flash Player 10 – I love speed
2. E4X String Parser

It’s one of the biggest traps you can fall into when developing a larger project where many objects are created and destroyed and created and so on… geeks call that kind of thing memory leaks I think :)

So if you can avoid it: avoid it! If you can’t.. may be this little thing can help:

package com.betabong.util
{
	import flash.utils.Dictionary;
 
	public class WeakMethodClosure {
 
		private var holder : Dictionary;
 
		function WeakMethodClosure( target : Object , method : Function , arguments : Array = null ) : void {
			holder = new Dictionary( true );
			holder[ target ] = 	new MethodStorage( method , arguments );
		}
 
		public function call() : Boolean {
			var args : Array;
			var f : Function;
			var cache : MethodStorage;
			for ( var obj : * in holder ) {
				cache = holder[obj] as MethodStorage;
				cache.method.apply( obj , cache.arguments );
				return true;
			}
			return false;
		}
	}
}
 
class MethodStorage {
	public var method : Function;
	public var arguments : Array;
	function MethodStorage( method : Function , arguments : Array = null ) : void {
		this.method = method;
		this.arguments = arguments;
	}
}

This actually uses one of the two mechanisms Flash provides to store objects by weak reference: the magic Dictionary. So instead of storing the function, you actually store the method storage or call method:

public function callAfterTime( f : Function , a : Array , t : uint ) : void {
this.storedfunction = ( new WeakMethodClosure( this , f , a ) ).call;
// and later:
storedfunction();

It might not be obvious where you’d wanna use this, but I guess if you’ve found your way to this blog entry, you might be able to find out if that’s of any value for you anyway ;)

Download WeakMethodClosure.as (zip)

Just be aware that this solution might not be superduper save in all cases: the Garbage Collector is a wayward beast – there’s no guarantee it’ll clean your dictionary on time. Still, in my experience, it’s save enough. It’s comparable to adding a weak event listener..

And hey, if you’ve got any better idea, I’d love to hear!

Update: Below a test application…

package
{
	import com.betabong.util.WeakMethodClosure;
	import flash.display.Sprite;
	import flash.events.TimerEvent;
	import flash.utils.Timer;
	import flash.utils.getTimer;
 
	public class WeakClosureTest extends Sprite
	{
		private var closure : WeakMethodClosure;
		private var timer : Timer;
 
		public function WeakClosureTest()
		{			
			testWeakClosure();
		}
 
		public function testWeakClosure() : void {
			var tempObject : Object = {
				test: testTrace
			}
			closure = new WeakMethodClosure( tempObject , tempObject.test );
			callWeakClosure();
 
			timer = new Timer( 50 );
			timer.addEventListener(TimerEvent.TIMER , callWeakClosure );
			timer.start();
		}
 
		public function callWeakClosure( event : TimerEvent = null ) : void {
			if ( !closure.call() ) {
				trace( "Closure failed" );
			}
		}
 
		public function testTrace() : void {
			trace("Called test");
		}
 
 
	}
}

Outputs:

Called test
Called test
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed
Closure failed

Related posts:

1. Metatunnel 1k Demo: AS vs. JS
2. Nested inner functions can be evil
3. MetaTunnel with Pixel Bender

There’s also one for Actionscript 3 (xpath-as3). But.. well… I wanted to go for some real speed! I like XPath a lot, but we now have native E4X selection in Actionscript 3, quite a different concept of node selection, and the conversion of XPath to E4X obviously results in quite a compromise in performance.

So all I need is a decent E4X parser. And hey, I found one! E4XParser from Digital Primates. It does its job really well, especially considering the very compact code it consists of. Thanks to some preparsing and caching, it’s also quite fast.

Still I thought I can do better :-) So I planted myself for a day (and a night) in front of my displays and hacked the hell out of it. The result is a little library which does pretty much the same thing as E4XParser, though pretty much more and a little faster too (15% to 50%). It’s about half as fast as the native E4X selection (once parsed). You can do nearly anything you can do with E4X. Use it like this:

import com.betabong.xml.e4x.E4X;
var result : XMLList = E4X.evaluate( xmllist , "author.( name.@last == 'Jobs' )" );
 
// E4X.evaluate( source : XMLList , expression : String ) : XMLList

 
If your source is XML, just do XMLList( xml ), if your result should be xml, do xml = result[0]

Test it here

Restrictions: You can’t use AND/OR in comparisions. So, this won’t go: author.( name.@first == ‘Steve’ && name.@last == ‘Jobs’ ) – though this is only a real limitations for OR. do this for AND: author.( name.@first == ‘Steve’ ).( name.@last == ‘Jobs’ ).

What you can do: Yes, you can do quite advanced stuff like author.( name.@first == name.@last ) or car.@rating.average() (one of the few proprietary functions I added). Or even

*..car.( @brand.toLowerCase() == 'volvo' ).( parent().( localName() == 'group' ).@rating > @rating )

– a weird example, I admit, but fancy, ain’t it? :-)

This is the first time ever I’m releasing part of my library as Open Source (MIT licence). As soon as I’ll find some time (and if I see any interest), I’m gonna put this into Google Code, so everybody can easily checkout and participate. Until then download it from here:

Download (zip 13kb)

No related posts.

Let’s take the simplest parsing as an example. Converting a CSS like string “width: 100%; maxwidth: 500; fontface: Bold; gap: 10;” to an object { width: “100%” , maxwidth: “500″ , fontface: “Bold” , gap: “10″ }

Compare these two methods to achieve this:

public static function simpleStringToObject_fast( input : String , o : Object = null ) : Object {
	if ( input == null || !Boolean(input) ) return o;
	if ( o == null ) o = {};
 
	var cursor : int = 0;
	var found_key : String;
	var index : int;
	while ( true ) {
 
		if ( found_key == null ) {
			index = input.indexOf( ':' , cursor );
			if ( index >= 0 ) {
				while( input.charAt(index-1) == ' ' ) index--;
				found_key = input.substring( cursor , index );
				cursor = index + 1;
				while( input.charAt(cursor) == ' ' ) cursor++;
			} else {
				break;
			}
		} else {
			index = input.indexOf( ';' , cursor );
			if ( index >= 0 ) {
				while( input.charAt(index-1) == ' ' ) index--;
				o[ found_key ] = input.substring( cursor , index );
				found_key = null;
				cursor = index + 1;
				while( input.charAt(cursor) == ' ' ) cursor++;
			} else {
				o[ found_key ] = input.substr( cursor );
				break;
			}
		}
	}
 
	return o;
}
 
public static function simpleStringToObject_slow( input : String , o : Object = null ) : Object {
	if ( input == null || !Boolean(input) ) return o;
	if ( o == null ) o = {};
 
	input = input.split(' ').join(''); // remove spaces
	var a : Array = input.split(';');
	var pair : Array;
	for ( var i:int=0 ; i 1 ) {
			o[ pair[0] ] = pair[1];
		}
	}
	return o;
}

The first one might look like some crap lacking any elegance and style, but it’s twice as fast as the second one! And hey, if you do this a million times, it can be 5 seconds instead of 10. Might not seem to be that much, but I think it’s certainly worth the extra coding :-)

(Update: I haven’t mentioned RegExp here, though it definitely deserves a mention. Thing is that Regular Expressions are just damned slow in Flash compared to String manipulations, especially with longer strings. This is kind of surprising because it’s implemented natively. But the implementation obviously isn’t too good performance wise. I hope to see improvements with F10, though haven’t read anything about it yet – and also haven’t tested yet with beta. This little article confirms my experience with RegExp: With a file of 10,000 lines, the string version is still instantaneous, but the regular expression version takes about five seconds. Even with 180,000 lines, the string version is immediate. We gave up on the regular expression version after over five minutes..)

Related posts:

1. Binary Fun – Bits in Bed with Actionscript
2. Nested inner functions can be evil
3. Metatunnel 1k Demo: AS vs. JS

Visual Speed

While Flash Player 9 or more precisely the underlying actionscript 3 interpreter have lead to great speed enhancements on the logic side (meaning much faster data manipulation), it’s still rather slow when it comes to rendering. In a project I’m currently working on there’s quite some stuff going on here, but when I check speed performance with the Flex Profiler, most time is spent for rendering. Significant time. And it’s not some freaky Papervision 3D stuff ;)

So with Flash Player 10 this will be better. Adobe finally improves rendering speed, and the good thing is that every exisiting project (at least >= f9, all?) will profit without republishing. I had done some little tests with a beta, and though I don’t remember the results, they looked quite promising (even on a Mac). I’m not sure though when they’re gonna use the graphic card to speed up – I don’t think it made use of that in my tests. Most people mention the hardware acceleration (some rendering tasks can be offloaded to your graphics card), but in my understanding the overhauled the entire rendering system, so everything and everyone should profit.

Vectors

Vectors are simply typed Arrays. A limited array. Ha! why would we want that?! For two reasons:

• Developers have more feedback while coding: the IDE will be able to tell you when you put some wrong stuff in an array – as for now you could mix whatever data in an array, the compiler wouldn’t complain. But above all, the IDE will be able to provide you help while typing: no more casting all the time to get to the methods of an array item!
• Speed. A typed array can be much faster. And it will be.

It looks like this:

var vector:Vector.<int> = new Vector.<int>();

I’m not too much a fan of the definition syntax – something like this would look better, I think:

var vector:Vector::int = new Vector::int();
// or
var vector:int[] = new int[]();

But I haven’t thought of it much. Adobe certainly thought about it. Or some ECMA comittee. (Copy-pasting in WordPress HTML mode also won’t work anymore.. pfff) Also I personally would have preferred not to have another class for that (just enhance the good old Array), but there are reasons for that too (backward compatibility, I guess, I can sing a song about that too). So anyway, I’m eagerly looking forward to use that stuff. I just love speed.

More infos here: Using Vectors in ActionScript 3 and Flash Player 10

Update: Ok, I can’t make to appear the new Vector syntax in WordPress. Now I like the syntax even less ;-)

Related posts:

1. Flash Debug Speed
2. Alchemy – THE cool new flash feature!

<?xml version="1.0" encoding="utf-8"?>
<mx:Application xmlns:mx="http://www.adobe.com/2006/mxml" layout="absolute" creationComplete="create()" click="destroy()">
	<mx:Script>
		<![CDATA[
			private var dispatcher : Sprite;
 
			private function create() : void {
				dispatcher = new Sprite();
				// listen to inner function
				dispatcher.addEventListener( Event.ENTER_FRAME , function( e : Event ) : void { trace( e.type ); } );
			}
 
			private function destroy() : void {
				// this should free created dispatcher from memory
				dispatcher = null;
				// what we would need is:
				// dispatcher.destroy() or dispatcher.removeEventListener( Event.ENTER_FRAME )
				// or dispatcher.removeAllEventListeners()
			}
		]]>
	</mx:Script>
</mx:Application>

The dispatcher will just live on. The mean thing is that in method «destroy» we don’t even have a chance to get rid of the sprite anymore – it just lives on and on (and dispatches its event every time it enters a frame to the listener function.) So 3 objects will just not be freed in memory because of our inner function listener: 1. the dispatcher object, 2. the listener function (the inner function) and 3. the creating object, because the listener function (or its method closure) is bound to the creating method create() scope.

We also can’t use the addEventListener’s flag to use weak references, because the listener method wouldn’t have any other references and would «dye» instantely. So what you should do right now is something like that:

<?xml version="1.0" encoding="utf-8"?>
<mx:Application xmlns:mx="http://www.adobe.com/2006/mxml" layout="absolute" creationComplete="create()" click="destroy()">
	<mx:Script>
		<![CDATA[
			private var dispatcher : Sprite;
 
			private function create() : void {
				dispatcher = new Sprite();
				dispatcher.addEventListener( Event.ENTER_FRAME , handler );
			}
 
			private function destroy() : void {
				dispatcher.removeEventListener( Event.ENTER_FRAME , handler );
				dispatcher = null;
			}
 
			private function handler( event : Event ) : void {
				trace( e.type );
			}
		]]>
	</mx:Script>
</mx:Application>

That’s clean code. Good bye dispatcher, good bye sprite! But sometimes it can be quite difficult to maintain this. I personally would wish that actionscript objects, especially the EventDispatcher class, would provide a destroy() method. Calling it gets rid of all event listeners and also would unlisten to everything it has listened to. And for DisplayObjects it would also perform other operations like removing itself from its parent and destroying all children.

I’d also like to see the following method like this in EventDispatcher:

EventDispatcher.removeListener( type : String = null , listener : Function = null , useCapture : Boolean = false );
 
// so this would remove all listeners
EventDispatcher.removeListener();
 
// and this all listeners to enter frame
EventDispatcher.removeListener( Event.ENTER_FRAME );

I get around many issues by implementing something like a IEventListener interface:

interface IEventListener {
function listenTo( dispatcher : IEventDispatcher , type : String , listener : Function , useCapture : Boolean = false ) : void;
functon unlistenTo( dispatcher : IEventDispatcher , type : String , listener : Function , useCapture : Boolean = false ) : void;
function destroy() : void; // will remove myself as event listener from all registered dispatchers
}

Way to go :-)

Related posts:

1. Weak Method Closure

Still you can struggle over unexpected behaviour. First of all there is no guarantee that objects are garbaged (well it’s probably not a memory thing.. more about unreferencing, what the heck do I know?!) as soon as there aren’t any references left. But, in my experience, weak references work quite good.. I didn’t have to worry about them living for some more seconds or so. But. But! They’ll live on at least until next rendering step, meaning until actionscript execution is done.

I struggled over this with a validation manager, that keeps tracks of unvalidated objects in a weak dictionary. So when object unvalidates, it tells so to the validator who will validate before rendering update. Now let’s assume the object shall be destroyed in the mean time – it will unlisten to dispatchers, it will unlink sprites and so on.. but may be it’s already registered in the validator. So it’ll stay alive to fullfill next validation processing. (Which led into uncaught errors due to null references in the destroyed object in my example.)

I could of course unregister at the validator at destroy. But.. hmm.. I want my engine to be as fast and robust as possible and unregistering can be difficult and time consuming (especially when 1000 objects will be destroyed). So what I’ve done is, I set a flag ‘destroyed’ when destruction is complete, and on each validation I’ll check for that flag:

public function validateDisplay() : void {
  if ( _render_dirty && !destroyed ) {
    render();
  }
}

Checking against a boolean flag is fast as hell, so that’s cool. And also, it won’t happen often (it was really a rare exceptional case actually). So besides weak references in AS3 we still have to be careful as hell about memory leaks. By the way, the profiler in Flex Pro is just fucking great :-)

On a side note: Any one any info on MethodClosure objects lying around in memory?? I know what they are used for, but it seems they live and live and live….

No related posts.

Finally I was forced to rethink this strategy because I wanted to let other people work on a project. As many actionscript projects, this one also heavily relies on external libraries, most of them hosted by Google Code, another one being my own (closed) lib. So should I tell people: «hey, just svn checkout from here and there and then relink the libs in the project properties but don’t check in the updated settings for christ’s sake!»No. Too shaky, really.

Fortunately, SVN provides a fucking cool way to solve this issue: svn:externals. Ha! This is so awesome, you won’t believe it ;-) Okay, here are the steps to follow (on a mac at least):

Create a flex project:

• project
  • html-template
  • src

Add a new directory:

• project
  • html-template
  • src
  • externals

Now put this on your svn, either by svn add or by putting in the repos and checking out again (well.. just as you did). If you’ve already got bin-debug or bin-release directories, just remove them on the repos and svn update (they are created with each publish in flex).

Now let’s say our project resides in ~/Documents/projects/project (path to project). Do the following in Terminal:

cd ~/Documents/projects/project/externals
svn svn propedit svn:externals .

This should open an editor (in my case I’ve defined textmate as my editor by doing:)

export EDITOR='mate -w'

Enter the following (example libraries):

caurinatweener http://tweener.googlecode.com/svn/trunk/as3/
as3corelib http://as3corelib.googlecode.com/svn/

You can also add a specific release:

caurinatweener -r23 http://tweener.googlecode.com/svn/trunk/as3/

(in TextMate save and close, if you use vim you know what to do anyway)

Now comes the magic part:

svn up

This will checkout all defined libraries into externals/xyz!! you do svn commit and now every time somebody checks out your project (or your externals directory) all the libs will be checked out automatically! I think this is just fucking awesome! :-)

Now in Eclipse all paths to your external libraries will be relative to the project (of course you have to add the libs there or change the paths – you should know how to do that ;))

Now all we need to is ignoring bin-release and bin-debug. We don’t want these directories to be in our svn repos:

cd ~/Documents/projects/project
svn svn propedit svn:ignore .

enter a list of directories or files to be ignored. in our case:

bin-debug
bin-release

That’s it! Thanks, svn!

Related posts:

1. Test Movie from Flex to Flash (easy way)
2. Develop in Flex Builder, publish in Flash IDE (Mac OS X)