304370 - (flash-badmatch) crash if flash is loaded and xserver (xorg) uses the composite extension and has the color depth set to 16 bit

Status: RESOLVED WORKSFORME

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

Description

[Dominik Dresel]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.10) Gecko/20050724 Firefox/1.0.6
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.10) Gecko/20050724 Firefox/1.0.6

If you display a webpage with a flash animation in it (for example
http://www.stadtplandienst.de) Firefox will crash. 

This crash only appears when the following things are true:
XServer uses the composite extension und has the color-depth set to 16 bit.
If the composite extension is switched off or the color-depth is set to 24 bit,
the problem doesn´t occour.

If the flash-plugin itself is renamed or deleted, everything works only the
flash is not shown any more ("Click here to download the plugin").

Reproducible: Always

Steps to Reproduce:
1. Set x-server to use the composite extension and set the color depth to 16 bit
2. Start firefox, goto a webpage with flash e.g. www.stadtplandienst.de
3. Crash
Actual Results:  
In the console, the following message is shown and the programm will die with
all remaining processes.

Errormessage:
The program 'Gecko' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadMatch (invalid parameter attributes)'.
  (Details: serial 164 error_code 8 request_code 147 minor_code 3)
  (Note to programmers: normally, X errors are reported asynchronously;
   that is, you will receive the error a while after causing it.
   To debug your program, run it with the --sync command line
   option to change this behavior. You can then get a meaningful
   backtrace from your debugger if you break on the gdk_x_error() function.)
firefox-bin exited with non-zero status (1)

Expected Results:  
Not crash und show flash correct

Comment 1

[Dominik Dresel]

Attachment: Buildconfig of firefox

Comment 2

[Dominik Dresel]

Attachment: Plugins installed in firefox

Comment 3

[Dominik Dresel]

Attachment: XOrg-configuration-file

Comment 4

[hamacker]

Yes, the same problem.
My configuration is Pentium4, 512M RAM, Debian sid, Xorg 6.8.2 :
X Window System Version 6.8.2 (Debian 6.8.2.dfsg.1-5 20050810052028
[e-mail])
Release Date: 9 February 2005
X Protocol Version 11, Revision 0, Release 6.8.2
Build Operating System: Linux 2.6.12-1-686 i686 [ELF]
Current Operating System: Linux toshiba 2.6.12-1-686 #1 Tue Aug 9 13:00:08 UTC
2005 i686
Build Date: 10 August 2005
        Before reporting problems, check http://wiki.X.Org
        to make sure that you have the latest version.
Module Loader present
OS Kernel: Linux version 2.6.12-1-686 (buildd@ninsei) (gcc version 4.0.2
20050806 (prerelease) (Debian 4.0.1-4)) #1 Tue Aug 9 13:00:08 UTC 2005 T

Its was test in deekpark 2 and snapshot 20050815.
Flash version 7 (lastest).

Konqueror use the same plugin but does not crash and runs well in xorg.

If XFree (not xorg) everthing is OK.

Comment 5

[hamacker]

To workaround this problem :
export XLIB_SKIP_ARGB_VISUALS=1
firefox&

in terminal.

Comment 6

[Adam Guthrie]

Please provide a talkback ID for the crash. http://kb.mozillazine.org/Talkback

Comment 7

[hamacker]

In my first installation I turn off the agent talkback, how can I turn on, again ?
(using deeppark)

Comment 8

[Tom Chiverton]

This bug is happening to me also, according to the Xorg guys @
https://bugs.freedesktop.org/show_bug.cgi?id=1336 this is a bug in 'flash plugin'.

This isn't unconfirmed, as a number of people have this problem, and have had it
for some time (just look on google). It's still there in the 1.5 beta
(Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8b4) Gecko/20050908 Firefox/1.4)n too.

Is there anything that can be done to go forward with this ? I'd include a
talkback ID, but (on this linux box) it doesn't appear to generate one.

Comment 9

[Kevin Brosnan]

*** Bug 308024 has been marked as a duplicate of this bug. ***

Comment 10

[Kevin Brosnan]

*** Bug 258497 has been marked as a duplicate of this bug. ***

Comment 11

[Kevin Brosnan]

*** Bug 265426 has been marked as a duplicate of this bug. ***

Comment 12

[Kevin Brosnan]

*** Bug 268060 has been marked as a duplicate of this bug. ***

Comment 13

[Kevin Brosnan]

*** Bug 279837 has been marked as a duplicate of this bug. ***

Comment 14

[Kevin Brosnan]

*** Bug 285008 has been marked as a duplicate of this bug. ***

Comment 15

[Kevin Brosnan]

*** Bug 309624 has been marked as a duplicate of this bug. ***

Comment 16

[Kevin Brosnan]

I am going to leave this unconfirmed Gavin said that this is likely an external
problem.

Comment 17

[Adam Guthrie]

For all of you who are seeing this, do you see this in the latest version of Flash? http://plugindoc.mozdev.org/linux.html#Flash

Comment 18

[Jim Bray]

 The plugin site referred to in the last comment lists the current version
of the flash plugin as 7.0r25, which I had, but the version one will get
if one pulls the thing down is 7.0 r61. I installed this and tested. The bug
is still present in Firefox 1.5-rc3 (and/or flash 7.0r61, and/or the xorg
server...).

Comment 19

[Adam Guthrie]

So I was thinking... is there any reason why we can't add `export XLIB_SKIP_ARGB_VISUALS' to the Firefox mozilla.in file in order to stop this from happening? It sounds like Flash is abusing the "extra visuals". http://xorg.freedesktop.org/X11R6.8.0/doc/RELNOTES5.html

Comment 20

[Jim Bray]

 It's pretty clear this is a flash problem. Hopefully someone has told them about
it.
  As a wishlist item: it would be nice if plugins could be encapsulated somehow
so that instead of the browser crashing, the plugin crashes and is disabled. It
looks like Firefox is getting an X error and just throwing in the towel. A little
more robustness and perseverance would be a good thing if possible.

Comment 21

[Kevin Brosnan]

Jim the last bit is Bug 156493 'Browser should tolerate plug-in (plugin) malfunctions, like with a separate (own) process' or bug 230017 'RFE: Run plugins in a separate thread'.

Comment 22

[Adam Guthrie]

*** Bug 319548 has been marked as a duplicate of this bug. ***

Comment 23

[Kevin Brosnan]

*** Bug 346598 has been marked as a duplicate of this bug. ***

Comment 24

[Eugene Seppel]

*** Bug 353538 has been marked as a duplicate of this bug. ***

Comment 25

[Adam Guthrie]

*** Bug 355785 has been marked as a duplicate of this bug. ***

Comment 26

[timeless]

*** Bug 188302 has been marked as a duplicate of this bug. ***

Comment 27

[timeless]

*** Bug 189761 has been marked as a duplicate of this bug. ***

Comment 28

[timeless]

*** Bug 317209 has been marked as a duplicate of this bug. ***

Comment 29

[timeless]

*** Bug 344910 has been marked as a duplicate of this bug. ***

Comment 30

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Is this still a problem with the Flash 9 beta?  If so, it's probably worth reporting a bug to the Flash folks since they're actively working on the Linux plugin these days.  See http://labs.adobe.com/technologies/flashplayer9/ .

Comment 31

[landolini]

(In reply to comment #30)
> Is this still a problem with the Flash 9 beta?  If so, it's probably worth
> reporting a bug to the Flash folks since they're actively working on the Linux
> plugin these days.  See http://labs.adobe.com/technologies/flashplayer9/ .
> 
the problem is both with flash 7.x and 9 beta.

Comment 32

[Flummox]

Bug 356882 seems to be a duplicate (slightly different error message)
I have reported it to Adobe.

Comment 33

[Adam Guthrie]

*** Bug 356882 has been marked as a duplicate of this bug. ***

Comment 34

[landolini]

(In reply to comment #29)
> *** Bug 344910 has been marked as a duplicate of this bug. ***
> 
  
I've built Firefox with options you suggested but I've no more
informations ,but the same message

The program 'Gecko' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadMatch (invalid parameter attributes)'.
  (Details: serial 120 error_code 8 request_code 146 minor_code 3)
  (Note to programmers: normally, X errors are reported asynchronously;
   that is, you will receive the error a while after causing it.
   To debug your program, run it with the --sync command line
   option to change this behavior. You can then get a meaningful
   backtrace from your debugger if you break on the gdk_x_error() function.)

Comment 35

[Adam Guthrie]

*** Bug 357735 has been marked as a duplicate of this bug. ***

Comment 36

[Adam Guthrie]

*** Bug 358288 has been marked as a duplicate of this bug. ***

Comment 37

[Adam Guthrie]

jmott, you work on Flash 9 for Adobe, right? Do they know about this problem, or can you forward this bug to the appropriate people?

Comment 38

[Anders Conbere]

*** Bug 358407 has been marked as a duplicate of this bug. ***

Comment 39

[Vitaly Bordug]

*** Bug 358507 has been marked as a duplicate of this bug. ***

Comment 40

[teofilis]

*** Bug 358525 has been marked as a duplicate of this bug. ***

Comment 41

[Tomá¹ Kalu¾a]

I have two comments that are worth to mention:
 1. The very same plugin worked just perfectly in Firefox 1.5.0.7, at least in my case, under same settings of X.org
 2. I know that the problem is not on Your side, but I do agree with the opinion that the whole Firefox should not crash when there is a problem in the plugin. Because if I did not like Firefox so much, I would not spend several days trying to find out, where the problem is, but I would swith to Opera, where the plugin did not work, but the browser was running, or to old firefox 1.5.0.7 and I would abbandon 2.0 for good.

Comment 42

[Michael Shoshani]

I reported one of these duplicate bugs, and would like to offer a new development. My problems with this started when I upgraded my Ubuntu from Dapper to Edgy. When I was running Dapper I had flashplugin-mozilla installed; somehow in the upgrade that was removed and flashplugin-nonfree was installed instead. 

Tonight I removed flashplugin-nonfree and installed flashplugin-mozilla, and Firefox worked with Flash pages exactly the way it had before the upgrade. Firefox no longer crashes when it encounters embedded Flash material.

I have absolutely no idea what is different in the code between the two versions of flashplugin, but apparently it makes a difference to Firefox.

Comment 43

[Flummox]

It seems that this might now be fixed in the next build of flash 9.
For details (and workarounds) see the post from October 30th at http://www.kaourantin.net/

Comment 44

[timeless]

*** Bug 359929 has been marked as a duplicate of this bug. ***

Comment 45

[Kevin Brosnan]

*** Bug 361342 has been marked as a duplicate of this bug. ***

Comment 53

[John Vivirito]

This was fixed in Flash 9, can we please close this bug if noone can still reproduce this? 

Comment 54

[philbert]

yes i can repoduce it by doing following:

- start firefox

- go to youtube

- search for "crysis"

- click on the video titled "Crysis Physics - 3'000 barrel explosion made with SandBox 2"

- click on video titled "Crysis - Mass Physics" at the list right

- if firefox didnt crash then go back to the first video with "backspace" or the back-button and select the video "Crysis - Mass Physics" again

... now firefox should crash!

my specs: ubuntu hardy, firefox 3.0, flash-nonfree 9

Comment 55

[Wayne Mery (:wsmwk)]

does it fail with flash v10?

Comment 56

[Joe Olstad]

Originally, I had the same issue Dominik Dresel reported at the very top.  The issue reported here:

XServer uses the composite extension und has the color-depth set to 16 bit.
If the composite extension is switched off or the color-depth is set to 24 bit,
the problem doesn´t occour.

Why is this important?  Well, lots of people still have older hardware that they want to use, such as my old Toshiba notebook that still works good as long as it's in 16 bit color mode.  

In any case, the plugin fails in this configuration and crashes the browser.  Perhaps there is a way just to terminate the plugin without crashing the browser in this case (secondary issue)

Primary issue is the XServer settings vs compatibility with the plugin in 16 bit color mode.

Comment 57

[smadayag]

this should have been resolved in FP9.  closing>resolved>worksforme...

Comment 58

[BMO Automation]

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