Burn :D!

Many people prefer to have one input class that keeps track of what the user is doing (keypresses, mouse positions etc.) which can be polled when you need to know if the user pressed a certain key for example. It's pretty nifty and removes a bunch of eventListeners from your code, making it much more readable and clean.

You've got some stuff going so here some feedback:

I experienced lag when there where more than say 5-8 bullets flying. Not acceptable on any modern computer! Fix it by going over your bullet loops and being smart.

Nice space invader enemies!

After 7 levels I stopped playing but I thought it was maybe too easy. Well it might get harder later so no real critisism, just letting you know.

Oh, and at this stage it's definitly a ALPHA version. When you think the game is complete, and I mean COMPLETE, the game is in BETA and then people test the real deal and you make small tweaks.

Good luck!

Oops, that was a bit embarasing...

So ignore my previous mis-informative ramblings about the stage. Goes to show that you should check the documentation before, eeh, rambling.

stage.addChild(preloader); //adding the preloader

stage.removeChild(preloader); //removing preloader and starting game. This is the problem

You should not add displayObjects to the stage, because a) it won't work and b) the Stage object is not a part of the display tree. What you probably want to do is

this.addChild(preloader);

and the same for removeChild().

Regarding your original question:
The caller is the object on which the function was called. I your code 'stage' (a Stage object) and in mine 'this' (which extends MovieClip).

At 10/12/09 12:49 PM, LeechmasterB wrote: I know!! But noone seems to understand my post or even try to listen. Also you can dip it down to O(n log n) when checking stuff against eachother. Yet my solution avoids that completely.

I'm with ya :D.

The heat map is probably a good idea if you manage to tweak it nicely, since it's cheaper O(n) vs O(n log n) at least for large number of particles. Question is, how large numbers do you need before the heap map wins? Updating a heat map might require a quite large constant before the O(n) if the particles should have more than very local influence.

I'd love to see someone experiment with it, because I can't really be bothered with 3 exams next week :P.

At 9/22/09 04:46 AM, Draglino wrote:

:... I didnt get what i should do but it isnt working...

Lol :D

At 9/14/09 04:04 PM, Tree-SkyLark-BCE wrote: http://opensource.adobe.com/wiki/display /flexunit/FlexUnit

Lol, FlexUnit actually seems to provide exactly what I need. Have to check it out when I get home but it should contain all solutions.

At 9/15/09 03:05 AM, henke37 wrote: This sounds like a standard preloading issue in an unusual shape. You got classes and you want them loaded, but you have no author time instances of them. Set them to export by force, problem solved.

I'm not sure what you mean... I know how to export classes without having instances, but don't want to export stuff - that's clumsy. The point is to get something JUnit-like that can be used for all my projects. Therefore I don't want to recompile the program for every project.

Ideal use case:

- run program

- specify .as file
If specified class extends TestBase: runs tests and display results. Repeat.

- be happy

Blerk, seems like making an itsy testing tool in flash aint gonna work very smoothly.

Lol, I've been coding in python all summer - these kind of thing work out really nice in python. Although it's a hopeless language in many other ways.

Weell,

I'm using getDefinitionByName() right now and it works fine for loading for example flash.display.Sprite, but it won't work on my own classes unless they are in the compilation root folder. That is unless I import them manually at compile time. Which I could do, but would rather not since I would have to make separate testing files for ever project.

Could you load .as files with a Loader and parse them in some way?

Ok, this is a sort of advanced topic and I'm not even sure it can be done.

What I'm loosely doing is a testing program in flash, sort of like JUnit. The idea is that the user specifies a class name (ie. 'flash.display.Sprite') that extends my 'TestBase' class. This is tested by using assert type functions specified in 'TestBase'. The problem is that I need to import the class to be tested dynamically, and I don't know how to do it. Erhm...

Ideas?

At 9/3/09 11:52 AM, GustTheASGuy wrote: The OP used 'continuous' in a different sense, as you can see in the very first sentence:

As in opposed to discrete, mathematically speaking =P.

Moving objects around makes the graph search much harder because you can no longer precalculate the distances between the nodes. This is not a problem in a tilebased environment because the move cost is always 1 and you can only make move to a finite number of maximum other nodes (typically 4 or 8). ...Ok I'm slipping of topic with no particular point...

Crap!

I also meant to say that you select groups with the mouse and move with SHIFT + CLICK

Yo,

Interestingly enough, I've implemented A* in a continous environment with manually placed nodes. Check it out.

A* searches are done groupwise, and if you are familiar with the algorithm you can probably figure out the coloured lines.

At 6/22/09 05:55 PM, Defrag wrote: Thanks! The math that does the lighting isn't too hard, it just takes some creativity; Under the hood, the ship graphic has both a 'graphic' that stores color information and a 'material' that stores surface normal and reflectiveness information. Then at runtime, the light on each pixel is calculated as a function of the distance between the pixel and the light source, as well as the ship's material information for that pixel. When adding color from multiple light sources, if a pixel's color components ever overstep 255, the extra brightness is stored in a 'bloom map', which is subsequently blurred using a Gausian-esque algorithm, and added back on top of the rendered image.

...In case you were interested =P

I am :D.

Ah, that's pretty cool. But then the artist needs to produce these normal maps at the same time as the graphic... Well it looks good anyway so I suppose it can be worth the extra work. And the bloom effect is nice as well, I think actually that I recall someone asking about FFT's a while back. I've been going over them at Uni but it was some time ago so I couldn't really help out...

Pretty nice lighting effect! Looks grand, although it runs a liiittle chunky. The distortions pretty standard tier 3 effect, but you got sparks fizzling through my brain with the lightning. I can imagine that it runs heavy but I'm not really in the mood to figure out how :D.

At 4/26/09 08:24 AM, dELtaluca wrote:

At 4/26/09 08:18 AM, KaynSlamdyke wrote: Luca made me a simple one that just does it straight to Sprite.graphics

More specificaly, it renders a triangle into the graphics with a bitmap fill, taking the 3 screen coordinates for the triangle the bitmap texture, and 3 texture coordinates.

Ah, same as mine then. Although I took the idea from some other guy :D.

But if he ain't transforming the "floor" tiles (assumption), couldn't you make a simplification of 3d space and modell the "wall" polygons as having parallell upper and lower sides, *breath*, allowing the use a line-based distorter where every pixel line parallell to the upper and lower side is simply scaled. OK so I'm just presenting an idea from the top of my head, it might be slower or it might look ugly, I haven't tried.

And sorry for breaking out of topic Kayn :P. It's a cool engine.

At 6/11/09 06:28 PM, zuperxtreme wrote: From what I understand you're telling Flash that your Class is going to have something to do with that extended Class.

Say you wanted to make a custom event you'd have to extend that.

I think it's called inheritance?

I'm not too sure.

Lol, grasping for straws?

By having a class A extend class B, class A inherites all of class B's properties. So if my new class "MegaMovieClip" extends MovieClip, I can in fact treat MegaMovieClip as a MovieClip in addition to using the properties and functions I myself wrote into MegaMovieClip.

This is very useful all the time, especially when you code large projects (mmm, games).

At 6/11/09 12:13 AM, MasterOfObscurity wrote:

player.onEnterFrame = function() {
 	function checkground():Boolean {

Can you really have nested functions? I doubt it... place checkground() outside the onEnterFrame.

while(checkground()==false&&jumped==false) {

I don't remember in which order flash does logical statements, but to be sure nothing funky's going on, I'd change this to

while((checkground() == false) && (jumped == false)) {

By the way, I haven't used AS2 in a long time, but I think your code was sort of neat in a way I remember AS2 not to be, with properly typed vars and stuff. Nice!

One way of fixing this is to add a small script add the end of your animations saying for example

_parent.gotoAndStop("idle");

if you call the frame with your idle animation movieclip "idle". As a rule, don't code games in the timeline, but this is the one exception I can accept. Otherwise you can check what frame you are on with the currentFrame property.

At 5/28/09 08:13 PM, Pendragon92 wrote:

: 		while(e.target.hitTestPoint(char.x,Number(char.y - 1),true) != false){
: 			char.y -= 1;
: 		};

The only thing I could think of would be that the Number() function could be causing strange stuff to happen in the function call. This is only based on your problem description and nothing on the mechanics of AS. The Number() function is not needed anyway, so try with simply

while(e.target.hitTestPoint(char.x, char.y - 1, true)) {

and I found the '!= false' statement to be confusing... By convention I think you shouldn't write it. Use '== true' or the short form as seen above.

At 4/24/09 11:43 AM, KaynSlamdyke wrote: - The first line

MAC 10,0,12,36

- Average number you see in the second line for both High Quality (default), and Low Quality (Right click, select it from the list)

Both about 50 fps

- Whether there's a drop in visual quality on Low Quality

Barely, as someone mentioned now and then you got a black line of pixels between two tiles on low quality but I onl noticed it because I was looking.

- Vague specs of your computer, including browser.

2 GHz dual core, 2 gig ram, Mac Book Pro from last christmas

- What the last number does, and how fast it climbs. And if it ever drops.

started at about 39, and climbed to 54 and stayed there

- What you see, and what you think should be done with it.

Looks cool and I think this could be used for tones of games. Dungeon crawlers or platformers for example.

With thanks to Luca for getting me started on it again, pointing me in the right direction and for providing the distorter I'm using.

Yeah, what distorter are you using? A triangle based on or I suppose a line based would be better?

At 4/13/09 12:01 PM, GustTheASGuy wrote: Fickludd: BitmapData.copyPixels isn't blitting exactly, but yeah, the greatest gain is that the splatters would be flattened to a single image which Flash doesn't do on its own.

Agreed, but sometimes an explanation simplified to incorrectness is more useful that a correct one :D. And I felt lazy.

Weeell, just add the graphics to a suitable display hierarchy and align them to the position and rotation of their corresponding box2d shapes every turn. If you can use box2d that should prove pretty easy

At 4/13/09 10:23 AM, Paranoia wrote:

At 4/13/09 10:01 AM, Cia-Vamp wrote:

Instead of attaching splashes as sprites, draw them onto a bitmap. Then it shouldn't make any difference how many you've actually got, because it's still just the same number of pixels on the same bitmap.

This is absolutely true and is called blitting. You can find an extensive tutorial about it here.

Actually 3000 hittests is not horribly much and is probably not the big lagg causer. Often what causes lagg is lots of unoptimized vector graphics moving about on the screen, or having and creating lots of movieclips on stage. This can be helped by

a) Using Shapes instead of Movieclips where possible
b) Making the objects that are numerous simple or at least optimizing them (leading to fewer lines)

OR

c) Using exclusively bitmaps and thus using the blitting technique.

Nice work on the start screen! I like your style with the fighters too, if you need a coder for any future games feel free to contact me and I'll see if i've got time.

I've done both animation and programming for the games i've finished. But although I started making movies I now se myself more like a programmer who does graphics on the side.

But well, it's not like I excel at graphics =P.

Also,

Very often in flash lagg comes from the flash player miscalculating the time required for calculations each frame, which causes dropped framerate. This can be much improved with a optimized game loop.

At 8/21/08 12:04 PM, Cecemel wrote:

At 8/21/08 11:48 AM, Arby wrote: EPIC FINAL

fuck you fickludd, why can't you shoot at the closest bot >:(
Why ALWAYS me?

if you shoot at _root.bots[0], i demand a rematch :P

not really but, arby, why do they always shoot at my bots?
I'd like to know :/

i still pwn your asses 1 on 1 though :P

LOL, i said my bots were rushed :D. The only thing that's advanced is the dodging, the is just the quickest working solutions I could find, like shoot at the first bot in _roots.bots that's in another team. This is a really crappy approach. I don't recomend it.

On another note about my bots:
- They get hit alot because i made a bug in the avoid other bots part of the dodging, so they think they can stand on each other to some degree.
- They have NO sense of team. Really they are each identical and only take care of themselves and shoot at stuff. As we can see in many battles, healing a team mate that's about to die can be very good.

Eh, when I watched the final i died first and ZuiGe won, that's what happened for everyone, no?

Lol, that was more like I expected.

gg GuyWithHisComp.