Loader.unload()'s purpose is remove a reference to the loaded content as
part of the cleanup. Only, it's implementation doesn't seem to be thought
out at all. Logically, it'd probably be part of some clearing operation
triggered by a click handler or something putting Loader.unload() a ways
down the call stack. Yet in that case it throws an error saying the content
is not a child of the caller. WTF? In order to make it available for garbage
collection AND deal with the error you have to:
try { Loader.unload(); } catch (e:*) { }which is just f*cking stupid. Further, it's nice to be able to call Loader.close() if in fact the user is click happy but there's no damn way to see if content is actually still loading aside from checking on bytes. So you end up having to use the same hollow try,catch crap. I don't support FP10 yet so unloadAndStop() is a day late and a dollar short. On Sat, Nov 22, 2008 at 9:52 PM, Olivier Besson <[EMAIL PROTECTED]> wrote: > it seems that Loader.unload is not designed to unload in the "AS2 way". > It only removes a reference to the loaded content. > unloadAndStop() (flash player 10) may do the job.. but not sure if we can > rely on it, yet. > > Moreover, removing a movieClip on the timeline (by puting a blank keyframe, > or by removeChild() or unload(), etc..) cannot be considered as a safe > "remove" anymore. You have to design you own "removeMovieClip" stuff, but it > require a bit work if you want something clean and robust. > (see 2006 Grant Skinner articles about memory management, or Moock chapter > about unloading assets) > > AS3 is a really bad joke it seems. >> > > I wouldn't say that as3 in itself is bad or anything else: what works is > ok, but the BUGS are bad ;) > Very very bad, sometimes. > Specially those related to the timeline since they mostly show very poor > testing from Adobe (I'm still wondering if they have proper in-house alpha > testers !?). Adobe can't exclusively rely on "free" beta-testers focused on > pure actionscript and (seemingly) rarely using the timeline, or rarely using > both the timeline and as3 (but I by no means I would criticize those > beta-testers since they do it for free and during "free" time). > > > Joel Stransky a écrit : > > This is the first advancement since Flash's inception that I'd kill to go >> backwards. The absolute simplest of flash movies with dynamic content >> borderline nonsense. I don't think Loader.unload() even works at all. AS3 >> is >> a really bad joke it seems. >> >> On Sat, Nov 22, 2008 at 6:41 PM, Olivier Besson <[EMAIL PROTECTED]> >> wrote: >> >> >> >>> yep, >>> >>> I've thoroughly practiced as3 only for few weeks, and I can also mention: >>> - script in frame 1 not executing if the clip is created in a Mouse.CLICK >>> event >>> - play() or gotoAndPlay() displays the 1st frame twice (easily visible in >>> 1 >>> fps) >>> - visible not always working with instances set up in IDE >>> >>> I heard that some of these bugs have been fixed in CS4/ Flash player 10 >>> (a >>> player that break existing content, for example some PNG are not >>> displayed: >>> funny), but I also heard that CS4 IDE is more buggy than CS3. >>> >>> CS5 is the last chance for getting a decent graphic+programmer >>> environment, >>> I think... Some hopes ...but more probably Flash may become as crappy as >>> some other big, monopolistic, overly complex, and hard to maintain, >>> softwares (commonly used in graphic design). >>> >>> It really really sucks ;) >>> >>> >>> Juan Pablo Califano a écrit : >>> >>> It seems like the timeline and classes don't mix well in Flash IDE + AS >>> >>> >>>> 3.0. >>>> I've had lots of really anoying problems when mixing them -- even those >>>> hacks seem to work randomly; the same action sometimes ends up in a >>>> NullPointerException (or whatever it's called) but sometimes, it >>>> doesn't... >>>> And I've also come across other quirks, like masks that are lost when >>>> you >>>> move backwards in the timeline (but not when you move forward), changes >>>> in >>>> stacking of author time assets, even though each object is in its own >>>> layer, >>>> and some other random crap. >>>> >>>> I can really see how AS 3.0 is far superior to AS 2.0, and I've enjoyed >>>> it >>>> in any non-Flash (I mean Flash IDE) project. I've also always disliked >>>> the >>>> whole "everything you want to programatically control has to be a >>>> MovieClip" >>>> thing. But, at least in FP 9 (I read this has been fixed in FP 10, but >>>> looks >>>> like it'd be a bit too much not to support FP 9 for this), mixing the >>>> timeline with classes seems to be asking for trouble. >>>> >>>> Too bad, because as a programmer, I'm not really interested in doing the >>>> layout of visual objects; any designer would do it better and faster, >>>> and >>>> it >>>> just bores me. But it seems that you have to choose between coding >>>> everything in the timeline (sigh) or adding and removing (and >>>> positioning) >>>> stuff by code. It really sucks... >>>> >>>> >>>> Cheers >>>> Juan Pablo Califano >>>> >>>> >>>> 2008/11/22, Joel Stransky <[EMAIL PROTECTED]>: >>>> >>>> >>>> >>>> >>>>> Yeah I've read a bunch of hacks that just go against my sense of >>>>> cleanliness. I was just venting. The real question here is what causes >>>>> a >>>>> playhead to play automatically. >>>>> >>>>> On Fri, Nov 21, 2008 at 11:23 PM, jonathan howe < >>>>> [EMAIL PROTECTED] >>>>> >>>>> >>>>> >>>>> >>>>>> wrote: >>>>>> Hi, Joel, >>>>>> >>>>>> I'm taking a longshot guess that you had the same problem I had last >>>>>> week >>>>>> when the list helped me...? >>>>>> >>>>>> I had an instance on a keyframe of the stage and it (sometimes) wasn't >>>>>> available i.e. null when I gotoAndPlay() to the frame it first >>>>>> appears? >>>>>> >>>>>> >>>>>> >>>>>> >>>>> Did >>>>> >>>>> >>>>> >>>>> >>>>>> you find a solution? I have had good success with the CustomWait class >>>>>> >>>>>> >>>>>> >>>>>> >>>>> from >>>>> >>>>> >>>>> >>>>> >>>>>> >>>>>> >>>>> >>>>> http://www.lecrabe.net/blog/index.php?post/2008/08/05/41-as3-gotoandstop-gotoandplay-issue >>>>> >>>>> >>>>> >>>>> >>>>>> . >>>>>> It is a sucky bug. Yeah: I call it a bug, not a feature. >>>>>> -jonathan >>>>>> >>>>>> >>>>>> On Fri, Nov 21, 2008 at 4:37 PM, Joel Stransky < >>>>>> [EMAIL PROTECTED] >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> wrote: >>>>>>> Don't worry, I doubt it even fk'ing matters anyway. I can't >>>>>>> access a >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> stage >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> instance on frame 2 or higher without some psychotic hack I've yet to >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> find >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> on google. How the f*ck did macromedia expect any common flash >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> designers >>>>>> >>>>>> >>>>> >>>>> >>>>>> to >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> keep up? I don't want to write a custom class for each stage instance >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> just >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> to utilize ADDED_TO_STAGE. What good is that to code on the timeline >>>>>>> looking >>>>>>> for that object? I just want to write a basic flash site with dynamic >>>>>>> content and avoid showing two preloaders just to load the home page >>>>>>> by >>>>>>> default. Jesus fk'ing christ! Very frustrated atm. >>>>>>> >>>>>>> On Fri, Nov 21, 2008 at 2:57 PM, Joel Stransky < >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> [EMAIL PROTECTED] >>>>>> >>>>>> >>>>> >>>>> >>>>>> wrote: >>>>>> >>>>>> >>>>>>> So I'm writing the most basic of timeline preloaders and >>>>>>>> flash seems >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> to >>>>>>> >>>>>>> >>>>>> >>>>> >>>>>> have a mind of its own. >>>>>> >>>>>> >>>>>>> My test.fla has two frames. Frame 1 has a text field and a progress >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> bar >>>>>>> >>>>>>> >>>>>> >>>>> >>>>>> movieclip. Frame 2 has a very large image. There is no code what so >>>>>> >>>>>> >>>>>>> >>>>>>>> >>>>>>> over. >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>>> The document class is set to Test.as and is essentially: >>>>>>> >>>>>>> >>>>>>>> package >>>>>>>> { >>>>>>>> import flash.display.MovieClip; >>>>>>>> import flash.text.TextField; >>>>>>>> import flash.events.*; >>>>>>>> public class Test extends MovieClip >>>>>>>> { >>>>>>>> //stage instances >>>>>>>> public var info_txt:TextField; >>>>>>>> public var loadBar:MovieClip; >>>>>>>> public function Test() >>>>>>>> { >>>>>>>> this.loaderInfo.addEventListener(ProgressEvent.PROGRESS, >>>>>>>> onProgress); >>>>>>>> this.loaderInfo.addEventListener(Event.COMPLETE, >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> onComplete); >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>>> } >>>>>>> >>>>>>> >>>>>>>> private function onProgress(e:ProgressEvent):void >>>>>>>> { >>>>>>>> trace(e.bytesLoaded); >>>>>>>> //on progress updates >>>>>>>> } >>>>>>>> >>>>>>>> private function onComplete(e:Event):void >>>>>>>> { >>>>>>>> //clean up >>>>>>>> } >>>>>>>> } >>>>>>>> } >>>>>>>> >>>>>>>> This all works fine except for the playhead activity. In my sample >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> you'll >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>>> notice there are no stop(); commands so when do Control -> Test >>>>>>> >>>>>>> >>>>>>>> >>>>>>>> >>>>>>> Movie, >>>>>>> >>>>>>> >>>>>> >>>>> >>>>>> it >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> just loops over the timeline flickering between frames 1 and 2 as you >>>>>>> >>>>>>> >>>>>>>> >>>>>>>> >>>>>>> would >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>>> expect. Oddly enough however, when I hit ctrl+enter again (Simulate >>>>>>>> Download), it stops, outputs the traces and then proceeds to >>>>>>>> flicker. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> Wha!? >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>>> It gets weirder. Just add a stop(); as the first line in the >>>>>>>> document >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> class >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>>> constructor right? Now Test Movie just sits on frame one and never >>>>>>>> broadcasts any ProgressEvent's. Seems fine but now when I Simulate >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> Download, >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>>> it stops, outputs the traces... AND THEN PROCEEDS TO FLICKER! My >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> onComplete >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>>> does NOT tell the playhead to play() or gotoAndPlay() or anything. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> Who >>>>>>> >>>>>>> >>>>>> >>>>> >>>>>> the >>>>>> >>>>>> >>>>>>> >>>>>>> >>>>>>>> heck it telling the playhead to play? >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> --Joel Stransky >>>>>>>> stranskydesign.com >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> -- >>>>>>> --Joel Stransky >>>>>>> stranskydesign.com >>>>>>> _______________________________________________ >>>>>>> Flashcoders mailing list >>>>>>> [email protected] >>>>>>> http://chattyfig.figleaf.com/mailman/listinfo/flashcoders >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>> -- >>>>>> -jonathan howe >>>>>> _______________________________________________ >>>>>> Flashcoders mailing list >>>>>> [email protected] >>>>>> http://chattyfig.figleaf.com/mailman/listinfo/flashcoders >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>> -- >>>>> --Joel Stransky >>>>> stranskydesign.com >>>>> _______________________________________________ >>>>> Flashcoders mailing list >>>>> [email protected] >>>>> http://chattyfig.figleaf.com/mailman/listinfo/flashcoders >>>>> >>>>> >>>>> >>>>> >>>>> >>>> _______________________________________________ >>>> Flashcoders mailing list >>>> [email protected] >>>> http://chattyfig.figleaf.com/mailman/listinfo/flashcoders >>>> >>>> >>>> >>>> >>>> >>> -- >>> Olivier Besson (gludion) - (33 1) 44 64 78 99 >>> http://wwww.gludion.com >>> http://blog.gludion.com >>> >>> >>> _______________________________________________ >>> Flashcoders mailing list >>> [email protected] >>> http://chattyfig.figleaf.com/mailman/listinfo/flashcoders >>> >>> >>> >> >> >> >> >> > > > -- > Olivier Besson (gludion) - (33 1) 44 64 78 99 > http://wwww.gludion.com > http://blog.gludion.com > > _______________________________________________ > Flashcoders mailing list > [email protected] > http://chattyfig.figleaf.com/mailman/listinfo/flashcoders > -- --Joel Stransky stranskydesign.com _______________________________________________ Flashcoders mailing list [email protected] http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
