463460 - crash with flash player 9.0.151.0

Status: RESOLVED INVALID

(External Software Affecting Firefox Graveyard :: Flash (Adobe), defect)

Description

[Rolf Bode-Meyer]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1b2pre) Gecko/20081106 Minefield/3.1b2pre
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1b2pre) Gecko/20081106 Minefield/3.1b2pre

I've just upgraded to newest flash-plugin 9.0.151.0 and Firefox keeps crashing on various sites that use flash. One of them is adobes site which has a flash banner on the entry page.
I've tested Firefox nightlies 20081106 and 20080905 as well as SeaMonkey nightly 20080918 which crash hard in that they just get terminated. No report, no trace.

SeaMonkey 1.1.12 on the other hand shows an alert "shockwave flash plug-in performed an illegal operation. You are strongly advised to restart firefox." and keeps going.

I had no problem with flash player plug-in 9.0.124 and don't have with 10.0.12.36.

Firefox 3.0.*, Opera, Safari 3.1.2 aren't affected and work normally.

Reproducible: Always

Steps to Reproduce:
1. Install flash player plugin version 9.0.151 with Firefox 3.1b*.
2. Go to Adobe website.

Actual Results:  
Hard crash.

Expected Results:  
Don't crash.

Windows XP SP3 on Pentium Dual-Core.

Comment 1

[Matthias Versen [:Matti]]

The latest flash player is v10, Adobe released the security upgrade V9 R115 only for users of Win98/ME and OS X10.1-10.3 but those OS are not supported by FF3.X.

That means that you should use v10 with FF3.X and nobody will invest much time for this issue because V10 works for us and that should be the last release of the v9 plugin. I bet that the crash player itself crashes, a stacktrace will show it, please follow this instructions :
https://developer.mozilla.org/en/How_to_get_a_stacktrace_with_WinDbg

Comment 2

[Rolf Bode-Meyer]

Attachment: Stacktrace for crash with Flashplayer 9.0.151

I'm aware of that limitation for that flash release. But prior Flash 9 releases where for 2k and XP too and I don't think Adobe limited the 9.0.151 to 9x/ME while fixing bugs. So the crash surely isn't because I'm using it on XP.
Secondly, as I wrote the crash doesn't happen in FF 3.0 or other vendors browsers.

Yes, Mozilla might not be interested in fixing that particular crash but I must say I was surprised to see FF simply vanish on the flash crash while SM 1.1 simply caught it and told the user about the problem. That's the way it should be for FF 3.1 too (besides not to crash in the first place). So isn't this simply reproducable crash a chance to change it?

For any case, I've attached a stack trace from WinDbg.

Comment 3

[Scott K]

For what it's worth, I'm also encountering the same crash with SeaMonkey 1.1.13 (and previously with SeaMonkey 1.1.12) on Win 2K when using Flash Player 9.0.151.0.  On some sites, rather than crashing, the plugin simply doesn't display everything correctly.  Previous versions of Flash Player didn't result in these crashes.  

While the latest version of the Flash Player (10.0.12.36) doesn't cause crashes either, there are still quite a few web sites that aren't working correctly with the it.  In order to access those web sites with SeaMonkey, an old version of the Flash Player plugin (with the security holes) has to be used.  Or the other alternative is to use FireFox 2.x with 9.0.151.0 Flash Player.  I haven't encountered a crash with that combination yet.  But it still seems like someone might want to look into this.  If Adobe put out a version of the Flash Player plugin that crashes some browsers, it's entirely possible that some future version may also cause crashes.  Updating SeaMonkey (and FireFox) to handle whatever is going on now may prevent future plugins from causing similar crashes.

Comment 4

[Matthias Versen [:Matti]]

>Yes, Mozilla might not be interested in fixing that particular crash
fix is the wrong word, investigate is the right one.
Why do you think that Mozilla can fix the plugin itself ?
If plugin version X crashes but the same plugin but with version Y works then this is in most cases a bug in the plugin itself.

A plugin can always crash the browser because it runs in the same process.
Your stack trace shows that the flash player itself crashes, please report this to adobe.

Scott: Your description shows that this is a bug in the player and not in Gecko. The will be AFAIK no future releases of v9. Adobe could only a new release with a fix in their plugin.

We have a RFE bug for out of process plugins. Gecko might survive a crashing plugin if this gets implemented.

Comment 5

[webbertiger]

Regarding plugin crash, I believe one important thing Mozilla can do is at least avoid the crashed plugin from crashing the browser--Google Chrome already did it. If some plugin crashes, it can just simply pop up a message to let the user know something is wrong with the plugin, and the browser is still running. This can also help users/developers report and trace the bug.

The same philosophy also applies to any good OS. An application crash should not crash the OS. Otherwise it would be a disaster.

Comment 6

[Matthias Versen [:Matti]]

[e-mail]:
directly above your comment :
>We have a RFE bug for out of process plugins. Gecko might survive a crashing
>plugin if this gets implemented.
This RFE is old (several years) and the developers are working on it.

Reporter: Do you still crash with the latest flash player version ?

Comment 7

[webbertiger]

(In reply to comment #6)
> [e-mail]:
> directly above your comment :
> >We have a RFE bug for out of process plugins. Gecko might survive a crashing
> >plugin if this gets implemented.
> This RFE is old (several years) and the developers are working on it.
> 
> Reporter: Do you still crash with the latest flash player version ?

Yes. I'm using Firefox 3.5.5 64-bit version. The flash player plugin (10.0.42.34) is available at
http://labs.adobe.com/downloads/flashplayer10_64bit.html
which is alpha version for x86_64.

I'm pretty sure in most cases the crash is caused by the plugin, but it just should not also crash Firefox. Especially when Firefox provide a open environment for plugin developers, it should also have some mechanism to isolate unstable plugins.

Comment 8

[timeless]

please don't repeat yourself. if you do it again, i'll mark this bug as invalid.

what's odd is that historically we did handle crashing plugins on windows. i can't figure out what actually happened in this path.

Comment 9

[webbertiger]

Don't use this attitude to talk to any bug reporters. I'm just trying to help Firefox to improve. A good engineer should know that a constructive suggestion is precious to the product.

Comment 10

[timeless]

thanks. i'll just wait for the other bug to be fixed.

Comment 11

[BMO Automation]

Version and milestone values are being reset to defaults as part of product refactoring.