Flash CS4: Movie goes blank while testing

[jacdx5050]

While testing the movie within Flash, the Test Movie window randomly goes blank, showing the stage color. The movie is still running, there's no error, it's just blank. When I resize the Test Movie window, everything reappears... Until a few seconds later, when it snaps back to blank... so I resize, it reappears, it disappears, I resize, it reappears, it disappears... over and over. Resizing even just one pixel is enough to get it to come back. Sometimes it happens every few seconds. It's enough to drive a person nuts.

This might be related to another issue: If I Test Movie with the Bandwidth Profiler on, Flash CS4 doesn't resize the window correctly to fit my entire movie in it, resulting in a big blank space underneath the Bandwidth Profiler, and that same amount of space being cutoff the bottom of my movie. I haven't seen anything like this in 10yrs of Flash development.

This is a Mac Pro, 8-core. The graphics card is ATI Radeon HD XT 2600. 6GB RAM. 10.5.3-10.5.7

The same FLA works without any issues on CS4/XP. The FLA saved to CS3 works without any issues on CS3/OSX and CS3/XP (same machine).

I have all but given up on CS4 for OSX. Between the crappy framerates, the crashing, and not being able to TEST MOVIE, I just avoid CS4 on OSX and use CS3 on OSX or CS4 on XP.

[Guest]

Hi,

Yes, Adobe is reading this page (please see all my posts).

If anyone experiencing this problem on a Windows machine can attach or send a FLA exhibiting this same issue (test movie goes blank until you resize the window), it would be helpful towards reproducing the bug on those machines. We have not seen this yet with the documents we have on the Windows end. Also please let us know all the details about your computer so we can check the file on a similar machine. We cannot necessarily fix bugs until we can reproduce them internally with a file or clear steps so we can create the problematic file.

So far the similarity between the files I have are: bitmaps are involved (created or imported) and if imported have transparency, there are either many of them or they are large, and the bitmaps change in some way (for example, animation or interactivity). All three are needed to reproduce. If your setup doesn't involve this, please attach a file that exhibits your problem so we can make sure it's tested when regressing a fixed version of Flash.

I don't think crashing is the same bug though - that sounds like something else is happening (not that it isn't valid or related of course, just something else going on).

[jacdx5050]

I haven't experienced this on XP (or 7 for that matter). I have switched to Windows 7 for all Flash CS4 development as this bug and the deplorable comparative performance of Flash on my 8-core Mac Pro vs my Win7 laptop have made it mandatory, unfortunately. The projects that I work on are all complicated (1-2 FLAs each and 50+ classes)...I haven't done any simple tests.

--J

[pyrografix]

I've also noticed this error with Adobe's TLF (Text Layout Framework) component. Specifically, the Window SWF that controls it. When you scroll the panel, it turns white. When you resize it, it goes back to normal. Just like the problems we see when we test our movies!

Hopefully that will narrow it down for you!

[Guest]

These are vary simple files and use only the flvplayback or  flvplayback 2.5 with AS3 to attach the skin and the video.

Most crashes occur when hard coding a web link to either in the skin or flv.  So local references work fine when referring to files with in the root folder structure appear. However this is not a solution for me.  I need hard coded links.

Crash behavior – testing the swf in the flash authoring environment
the fatal white screen of death appears.
If I close the window, moments later the app freezes, or
If I don’t close the window, a bit more time later the app freezes
Then the crash report.

Once Flash crashes – I usually have to restart my HP however, if I wait long enough I can restart the app.  Looking through the task list I can find an obvious app to force quit.

One final note, this is by far the simplest case in which this issue occurs.  Typically I work on projects that have 1000s of lines of code.  But simple is good, and I can’t share the other apps with you.

I will seek out your solutions and pray they fit for me.  For I did reinstalled Flash and this didn’t correct the problem.

My next thought is to remove the CS4 Masters Collection, flash player  and delete the registry keys.  But I really don’t want to do that.

[Guest]

yet another attempt to attach files.

[Guest]

Attempted to attach the source files and was unaware that the post went through and the attachment failed.

Dang!  Thwarted by my own network. 

[aj-runtimelogic]

I just read through this entire thread and wanted to get involved in the customer-side debugging effort.  I've been experiencing this issue too under Flash CS4 with a similar hardware configuration to those mentioned previously:

  Model Name: Mac Pro

  Model Identifier: MacPro3,1

  Processor Name: Quad-Core Intel Xeon

  Processor Speed: 2.8 GHz

  Number Of Processors: 2

  Total Number Of Cores: 8

  L2 Cache (per processor): 12 MB

  Memory: 6 GB

  NVIDIA GeForce 8800 GT / 512MB VRAM

I have seen this running under both Mac OS X 10.5 and after recently upgrading to Snow Leopard (10.6).  My Flash CS4 is patch to 10.0.2.

I agree that any crashing is a secondary issue to the seemingly random blanking of the display when testing the movie.  I actually haven't had any crashing under Flash CS4 like some people report so I think we should keep this thread focused on helping Adobe determine repeatable steps to reproduce so that one of their engineers can figure out where to set breakpoints and catch the failure state in a debugger.

Here are some observations on causes that I hope will help:

The project that I've seen this behavior in occasionally when testing just starting seeing this behavior constantly.  I went from having a static vector only Sprite in the background to an animated MovieClip involving moving a vector with an alpha gradient within a mask.  This by itself doesn't cause a problem - I isolated this in a separate FLA and test movie works without ever triggering the bug.  There are Sprites with dynamically loaded images nested in them that animate over top of this background based on a Timer.  That still doesn't cause any blanking if you leave it alone and just let it animate.

What does cause consistent blanking is rolling the mouse over the nested image Sprite which has a MouseEvent.MOUSE_OVER handler that causes a scale tween (via TweenLite).  It doesn't blank immediately during the tween, but after that a rectangular section of these items seems to be having redraw issues: there is some white flicker where I'm assuming I'm seeing the stage color temporarily.  After some period of time the blanking will occur.  Rolling over the other nested image Sprites on stage that also have MouseEvent.MOUSE_OVER handlers but don't trigger a scale tween will not trigger the bug.

I turned on "view redraw regions" to see if I could find a correlation.  It's hard to tell since there is so much going on but to me it looks like when the alpha gradient passes below the mouse position, that is when the bug occurs.

This repros for me almost every time.  If I see the redraw flicker, the blanking bug will be triggered.  It seems like this has to do with the position of the mouse and some issue with drawing to screen with multiple layers involving transparency.

Please let me know how I can be of further assistance as this is a big time waster for sure and it seems like we are getting close to a fix here.  I'm not able to post the project up publicly, but I can provide it to you directly for debugging if you like.

[Guest]

Hi there!

I'm not sure if the problem I'm experiencing is the same as the above, but I suspect it is similar. I have CS4 running on Windows XP and I'm working on a year project for University in Flash so when this bug started appearing today, I got a little stressed up about it since the work needs to be done and I cannot continue with it because of this bug.

In my case I'm drawing the background for a side-scroller game, but the movieclip simply goes invisible on the stage while I'm editing it. I don't know why this only happened now, I've been drawing for days without problems. The movieclip isn't visible on the main timeline nor on its internal timeline. Even if I drag the movieclip into a new .fla from the library the same thing happens over and over again.

I cannot tell you what my computer's specs are off the top of my head at the moment, but I will find out and tell you.

Here's the problematic file.

Thank you and I really hope there is a solution to this problem cause I'm panicing!

Ans

[Guest]

Hi there!

I'm not sure if the problem I'm experiencing is the same as the above, but I suspect it is similar. I have CS4 running on Windows XP and I'm working on a year project for University in Flash so when this bug started appearing today, I got a little stressed up about it since the work needs to be done and I cannot continue with it because of this bug.

In my case I'm drawing the background for a side-scroller game, but the movieclip simply goes invisible on the stage while I'm editing it. I don't know why this only happened now, I've been drawing for days without problems. The movieclip isn't visible on the main timeline nor on its internal timeline. Even if I drag the movieclip into a new .fla from the library the same thing happens over and over again.

I cannot tell you what my computer's specs are off the top of my head at the moment, but I will find out and tell you.

I will attach a file in another post.

Thank you and I really hope there is a solution to this problem cause I'm panicing!

Ans

[pyrografix]

Those are not the same symptoms we are experiencing. But I have had that problem on a Mac too.

Basically it happens when you create a new movieClip. It will suddenly go invisible, but you can select it (even though the bounding box is way off.) You can even double click it to edit the symbol. The only "fix" I found is to save your file, shut down Flash and restart it.

It's annoying.

It's buggy...

It's consistant for Flash CS4.

[oldschoolmx]

I believe I experience the same problem. And it is occurring with ALL of my Flash projects in CS4. When I export movie the file will launch but almost always after testing for ten to thirty seconds....the whole screen will go blank (white). When I test the same file in a browser it works fine. There are no code errors.

I am on a Mac with 10.5.6 and CS4.

I could not agree more with a previous poster:

It's annoying.

It's buggy...

It's consistant for Flash CS4.

[Antifuzz]

I'm going to Flash On The Beach in Brighton next week. I'm going to try my best to corner some Adobe bods and badger them about this. I'm getting more and more annoyed by Adobe's lack of interest in sorting this out.

[pyrografix]

Lucky bastard! Although, I'm going on my honeymoon in the Mediteranean.  So I will have fun too!

Anyway, give 'em hell. This application is almost useless. It crashes all the time, and not to mention debugging is impossible in the IDE since it goes blank to the stage. The actionscript editor is beyond useless, because half the time the auto format deletes code. I thought I would like the new motion editor. But guess what? That's horrible because it doesn't behave as expected. PLUS it's honky, because it take a lot of time to sort through every porperty. BESIDES, the new interface sucks... period. It's a big waste of real estate, and is unresponsive. (Half the time you click on stuff expecting something to be selected or updated, and you have to reselect several times before the application actually does what you were expecting.

I feel they ruined an application that is very near and dear to me.

[Guest]

As noted a few times earlier, I can assure you there is interest in sorting it out. There are multiple teams looking at the problem now. And yes, I encourage you to speak to whomever you might run into and show them the problem with your feedback first hand (although I encourage you to tell them there is a bug filed and it has been reproduced internally, so they know who to talk to in the office).

[Antifuzz]

Hi All,

Well I'm back from Flash on the Beach which was really really great. I'd never been before and found it really inspiring all round.

I went to the "Adobe Town Hall" session and asked a question about this bug, adding that it was very seriously affecting my own use of Flash CS5, along with several others (mentioning this thread). The response I got was actually quite confusing and not all that satisfying.

I believe it was Richard Galvan who actually responded to my question, and his initial response was "have you downloaded the 10.0.2 update?". I said yes and that the issue was still there. He then claimed that they can't reproduce the bug internally at which point I'm afraid, Jen, I mentioned your name and the fact that you had told me you were able to reproduce it internally.

I said that I just wanted to know that they were committed to fixing bugs like this in the current release. He said that they were, but admitted at that point that he was unaware of the current status of this particular bug.

So there you have it.

[Guest]

Hi eyegas,

I'm sorry that the response at the conference was confusing - I know how frustrating that can be. I'm really glad that you brought it up, and glad that you mentioned my name. One possible reason that he hasn't seen it is because we've been working on the engineering side on the problem (the engineering managers are definitely aware of this, they're the ones I've been working with). Big team and a (ahem) number of bugs flying around, sometimes paths don't cross -- and that's why I was hoping it'd come up at the conference, it can mean a lot.

Anyway, the current status of this is that we're working with the player team (where this bug is located, a player part) to try and make something work out. This is not over/ignored, and although I can't make any promises - I'm not a developer making a fix or manager making the decisions, but we are trying to make something work for you guys. Sorry for being vague, but I think that's all I can say.

Oh yes, and one of the questions that came up during the meetings I've had on this issue, for this forum, is they're asking how the workaround of testing in the browser works/doesn't work in your workflow. Test movie is a simulation, and there can be issues with test movie syncing up with what's in the browser players. And for instance, sometimes players are updated before we can update the player in authoring (test movie). What are your general feelings about testing in the browser instead (and so forth- any feedback on that would be great).

Thanks!

Jen.

[stiv41]

Thanks Jen...

Personally, testing in the browser or standalone Flash Player is a big hit to my productivity. I use trace() extensively, and testing movies without having the output panel to the side takes away one of the best tools I have to keep an eye on exactly what's going on.

And, if I go through this process 100 times a day:

make a change - test movie - make a change - test movie...

And then change that to this process:

make a change - publish - switch to browser - empty cache - reload - switch to flash - make a change - publish - switch to browser - empty cache - reload...

then that's a big difference in the time it takes to get anything done!

[pyrografix]

I would have to agree with you on this one. It kind of defeats the purpose of having a preview. I realize it's not a true representation of the final application. But honestly, sometimes you just want to trace a variable or find out which object is returning null.

At my job, we are not allowed to install other software to expose the traces... like firebug. So I am stuck creating text fields and "tracing" that way.

Previeiwing in the browser is not good alternative as stiv41 has noted. The best alternative so far has been going back to CS3. Our entire team has already.

[Antifuzz]

Hi Jen,

Thanks for the info, it's good to have your updates that something is happening. When I asked the question at the conference there was a panel of six Adobe people, including Lee Brimelow and I think Mike Chambers and Mark Anders. I'm not sure who the others were.

Regarding your point about testing in the browser, I have to agree with the others above that whilst it's certainly something I do regularly, it's not really a replacement for the test movie window as I also use traces a lot for testing and debugging, besides being an extra finger-dance to switch apps etc each time you want to test something.

[oldschoolmx]

Jen you are asking on behalf of others, "What are your general feelings about testing in the browser instead..."

Ok...here is my *honest* *gut* reaction to that question.

What the f---? It just sounds beyond lame to me that someone within Adobe would even ask that question.

Preview is a CORE feature to flash. I'd compare it to how walking is a core feature of being human.

I'm sure there are some technical hurdles that we outsiders don't understand but from the outside it just makes Adobe look incompetent. I know it's frustrating internally but it just doesn't feel like this issue has been appropriately escalated to a "must-fix now" priority.

There are a LOT of people having this issue and it's HUGE.

[aj-runtimelogic]

Hi Jen,

I also wanted to respond to your question about testing in the browser as a temporary work-around.

For the way I use Flash this isn't a work around.  I rely extensively on the ability to set breakpoints and use the Flash debugger so having to run in the browser breaks the tight integration between the two.  There is remote debugging, but this would make for a very awkward and inefficient workflow.

I also agree completely with eyegas and use test movie so many times a day that having to run in the browser would really add up to a large waste of time.  In addition, during general development, running from the Flash IDE disables certain annoying security behaviors when running a project locally that is ultimately deployed to a web server.  When I do a test movie, I'm not focused on testing these deploy issues but just want to test the latest functionality I implemented.

Thanks.  I look forward to further updates as you guys work on this issue.

[Guest]

Thanks everyone for the responses and helpful input. I will pass this on to the person who requested this feedback.

[Guest]

As I read some of the comments above, debugging functionality seems to be a concern.  However, Flex tests it's swf's in the local and the debug break points and all debug features still work.  In fact I work on a highly secure network and Flex bug checking is far easier. 

I wouldn't see this work around as a problem at all - save the security sandbox issues.  (I think) that in the flash testing environment suspends most of the sandbox issues, but I am certainly not sure about that. 

But that's my take,

acs

[Antifuzz]

You're right but that's kind of beside the point. We don't all have a copy of Flex Builder and shouldn't have to fork out for another piece of software just because the one we paid for has such a bad bug.

Personally I've lately given up on the Flash IDE, (largely because of this bug), and bought myself a copy of FDT which I can thoroughly recommend. At least I didn't spend my money on more Adobe software.

[Guest]

Here is the problematic file.

Thank you once again!

Ans