firefox crashed [@g_type_check_instance] [@g_signal_emit_valist] [@g_signal_emit]

Bug #122726 reported by taro kawagishi
14
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Eclipse
Invalid
Medium
firefox (Ubuntu)
Invalid
High
Mozilla Bugs

Bug Description

Binary package hint: firefox

... firefox crashed

...after reaching 100% CPU use for about 10 seconds.

It doesn't reproduce even after opening all the URL's in different tabs exactly the same way as before.

Possible test case: (from https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/122726/comments/2)
I opened a page with a Java applet and seeing nothing interesting there
I went back to the previous page without Java applet.
(one of links at
http://services.alphaworks.ibm.com/manyeyes/browse/visualizations)
I always disable Java in Firefox because applets often crash it.
After viewing the applet with Mozilla, I went back to Firefox, then it
crashed.

Extracted from symbolized stacktrace:
(file: http://launchpadlibrarian.net/8220995/gdb.log)
...
#3 <signal handler called>
#4 g_type_check_instance () from /usr/lib/libgobject-2.0.so.0
#5 g_signal_emit_valist () from /usr/lib/libgobject-2.0.so.0
#6 g_signal_emit () from /usr/lib/libgobject-2.0.so.0
#7 gtk_widget_event_internal (widget=0xb6d92de8, event=0xbfcba8e8)
#0 __kernel_vsyscall ()
#1 poll () from /lib/tls/i686/cmov/libc.so.6
#2 PR_Poll (pds=0x811a428, npds=1, timeout=4294967295) at ptio.c:3877
#3 nsSocketTransportService::Poll (this=0x8119f40, interval=0xb5c14428)
#4 nsSocketTransportService::Run (this=0x8119f40)
...

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Under Ubuntu Edgy Eft, using GLib 2.12.4-0ubuntu1, eclipse 3.2 and 3.3M2 crashes after a few minutes of using.

The crash happens when compiling java projects. Disabling automatic build greatly reduce the problem but it still happens from time to time.

Under Ubuntu Dapper Drake, which uses GLib 2.10.3-0ubuntu1, the problem does not happen.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Created attachment 52493
JVM crash log

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Created attachment 52494
Another crash

This time using JDK 1.6.0.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Created attachment 52782
Same bug, but this time using GLib 2.10.3

What appears to be the same bug, but this time using GLib 2.10.3.

The bug is much harder to replicate using GLib 2.10.3 (Ubuntu Dapper Drake), but it seems to be there too.

Revision history for this message
In , Snorthov (snorthov) wrote :

NOTE: All crashes are happening when gtk_widget_show() is used to show a top level shell. Anybody got any ideas? Perhaps Cedric could try running a modified version of Shell.setVisible()?

What's all this about g_type_check_instance_cast? Is that a clue?

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Yes, if there is anything I can do to help diagnose the problem let me know. I am no SWT hacker so just let me know what you need me to do / provide.

Given the large Ubuntu user base, I suppose there will be many Eclipse users with that problem, especially since I have been able to replicate the problem on 2 different ubuntu workstations running different releases (6.06 and 6.10).

Revision history for this message
In , Snorthov (snorthov) wrote :

Bogdan, give him a hacked Shell.setVisible() that doesn't wait for all paints to have happened. Ask Silenio what this means.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Using Eclipse 3.3M4 under sun-jre 1.6.0, I can no longer reproduce the problem.

I have also upgraded to Ubuntu Feisty Fawn (dev version), which uses libgtop 2.14.8.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

I meant GLib 2.12.9. Don't know if it matters anyway.

Revision history for this message
In , Snorthov (snorthov) wrote :

Ok, WORKSFORME?

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Unfortunately I spoke too soon.

I had what appears to be the same crash again. It does happen a lot less often now. The crash happened while I was doing a 'relocate' operation using subclipse plugin.

The other crashes happened mainly while (auto-)building java projects. I haven't had these under eclipse 3.3M4 (yet).

See the jvm crash log 'Same crash using Eclipse 3.3M4' that I will attach.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Created attachment 60033
Same crash using Eclipse 3.3M4

Revision history for this message
In , Snorthov (snorthov) wrote :

Bogdan, get that man a hacked Shell.setVisible().

Revision history for this message
In , Gheorghe-ca (gheorghe-ca) wrote :

Created attachment 60082
Hacked SWT jar

Here is a replacement jar that contains the hacked Shell.setVisible (the change to Shell is to no longer wait for all of the paint events). To use this:

1. Extract a new version of Eclipse M4.
2. Place the attached jar file in [extracted eclipse dir]/eclipse/plugins -> you will be overwriting the existing jar
3. Start up Eclipse from a terminaal -> ./eclipse
4. If everything went OK you will see the message: "Hacked Shell.setVisible: not waiting for paint events" displayed in the terminal.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Ok, running the hacked swt jar and seeing the

Hacked Shell.setVisible: not waiting for paint events

debug output.

So far no crash after a few hours of use.

Revision history for this message
In , claudio (private-claudio) wrote :

Created attachment 63189
crash dump

Revision history for this message
In , claudio (private-claudio) wrote :

The above is a crash dump in gtk which I get with eclipse 3.3M6. Is this the same problem or a new one?

Revision history for this message
In , Snorthov (snorthov) wrote :

Seems differed (ie. OS.gtk_tree_store_remove()). SSQ, didn't we fix a GP in this area a while back?

Revision history for this message
In , Gheorghe-ca (gheorghe-ca) wrote :

Claudio, I think you are seeing Bug 167562.

Revision history for this message
In , Snorthov (snorthov) wrote :

Ok, the GP is still open.

Revision history for this message
In , Gheorghe-ca (gheorghe-ca) wrote :

BTW, Claudio - do you happen to have a set of reproducible steps for your crash?

Revision history for this message
In , Jeremy Volkman (jvolkman) wrote :

I had a similar crash and could reproduce it 100% by unchecking and checking a bundle in an OSGi launch configuration. I've fixed the problem by disabling the "GTK Accessibility Module". In Ubuntu Feisty Fawn (and perhaps earlier), this setting is in System->Preferences->Accessibility->Assistive Technology Preferences.

Revision history for this message
In , Gheorghe-ca (gheorghe-ca) wrote :

Interesting - I tried to reproduce this (Accessibility module on + checking/unchecking OSGi bundles in launcher) on RHEL5 but no luck.

Cedric, do you have the Accessibility module turned on?

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Yes, I do have the Accessibility Module turned on.

I have been using eclipse 3.3M5 (Build id: I20070222-0951) for a few weeks now and never had that crash.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Jeremy is right, checking / unchecking OSGi bundles in launcher config does reproduce the problem.

I had to play with the check boxes a little bit and then it crashed.

I will attache the crash dump. This is under 3.3M5.

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Created attachment 65399
Crash under 3.3M5 - checking/unchecking OSGi bundles

Revision history for this message
In , Jeremy Volkman (jvolkman) wrote :

(In reply to comment #24)
> Jeremy is right, checking / unchecking OSGi bundles in launcher config does
> reproduce the problem.
>
> I had to play with the check boxes a little bit and then it crashed.
>
> I will attache the crash dump. This is under 3.3M5.
>

Can you verify that it doesn't happen with the Accessibility module disabled?

Revision history for this message
In , Cedric Veilleux (cveilleux) wrote :

Yes, I confirm it does not happen with the Accessibility Module turned off.

Revision history for this message
Justin Payne (neolithium82) wrote : Re: Firefox crashed after reaching 100% CPU use for about 10 seconds.

Thank you for the bug report. Could you please try to obtain a backtrace by following the instructions on https://wiki.ubuntu.com/MozillaTeam/Bugs, or upload the crash report from /var/crash/.

Also, please answer these questions:
Which flash package do you have installed?
Which Java package do you have installed?
Which firefox extensions do you have installed?

This will greatly aid us in tracking down your problem.

Regards,
Justin

Changed in firefox:
assignee: nobody → mozilla-bugs
status: New → Incomplete
Revision history for this message
taro kawagishi (taro-kawagishi) wrote : Re: [Bug 122726] Re: Firefox crashed after reaching 100% CPU use for about 10 seconds.

Hi Justin,

I remember now what could have caused this crash.
I opened a page with a Java applet and seeing nothing interesting there
I went back to the previous page without Java applet.
(one of links at
http://services.alphaworks.ibm.com/manyeyes/browse/visualizations)
I always disable Java in Firefox because applets often crash it.
After viewing the applet with Mozilla, I went back to Firefox, then it
crashed.

The crash report is huge:
taro@ubuntu-taro:~$
ll /var/crash/_usr_lib_firefox_firefox-bin.2010.crash
-rw------- 1 taro exava 53214785 Jun 28
09:23 /var/crash/_usr_lib_firefox_firefox-bin.2010.crash
taro@ubuntu-taro:~$

Attached is the stack trace I got following your instruction.

External packages I use are:
Java from Sun jdk1.5.0_08 (I don't know how it affects Firefox when I
use it with "Enable Java" unchecked.)
Flash Player 9 for Linux Version 9.0.31.0
Firefox addons installed are
 DOM Inspector 1.8.1.3
 Event Spy 0.2.2
 Firebug 1.05
 Html Validator 0.8.3.9
 Inspect Context 0.3
 InspectThis 0.2.9
 User Agent Swithcher 0.6.10
 XML Developer Toolbar 0.2

Best regards,
Taro

-------- Original Message --------
From: Justin Payne <email address hidden>
Reply-To: Bug 122726 <email address hidden>
To: <email address hidden>
Subject: [Bug 122726] Re: Firefox crashed after reaching 100% CPU use
for about 10 seconds.
Date: Thu, 28 Jun 2007 01:27:02 -0000

Thank you for the bug report. Could you please try to obtain a backtrace
by following the instructions on
https://wiki.ubuntu.com/MozillaTeam/Bugs, or upload the crash report
from /var/crash/.

Also, please answer these questions:
Which flash package do you have installed?
Which Java package do you have installed?
Which firefox extensions do you have installed?

This will greatly aid us in tracking down your problem.

Regards,
Justin

** Summary changed:

- It just crashed after reaching 100% CPU use for about 10 seconds.
+ Firefox crashed after reaching 100% CPU use for about 10 seconds.

** Changed in: firefox (Ubuntu)
     Assignee: (unassigned) => Mozilla Bugs
       Status: New => Incomplete

Revision history for this message
Hilario J. Montoliu (hjmf) (hmontoliu) wrote : Re: Firefox crashed after reaching 100% CPU use for about 10 seconds.

hmm, there seems not to be any attachment,

please, can you resubmit the attachment (either the full crash report, or the retrace you did) to [1] instead of directly attach it to the email.

Thank you in advance,

H. Montoliu

[1] https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/122726/+addcomment

Changed in firefox:
importance: Undecided → High
description: updated
Revision history for this message
taro kawagishi (taro-kawagishi) wrote : Re: firefox crashed

I remember now what could have caused this crash.
I opened a page with a Java applet and seeing nothing interesting there
I went back to the previous page without Java applet.
(one of links at
http://services.alphaworks.ibm.com/manyeyes/browse/visualizations)
I always disable Java in Firefox because applets often crash it.
After viewing the applet with Mozilla, I went back to Firefox, then it
crashed.

Attached is the stack trace I got following your instruction.

External packages I use are:
Java from Sun jdk1.5.0_08 (I don't know how it affects Firefox when I
use it with "Enable Java" unchecked.)
Flash Player 9 for Linux Version 9.0.31.0
Firefox addons installed are
        DOM Inspector 1.8.1.3
        Event Spy 0.2.2
        Firebug 1.05
        Html Validator 0.8.3.9
        Inspect Context 0.3
        InspectThis 0.2.9
        User Agent Swithcher 0.6.10
        XML Developer Toolbar 0.2

description: updated
Revision history for this message
Hilario J. Montoliu (hjmf) (hmontoliu) wrote :

Thank you for the bug report however this report lacks information we need to investigate it further. Of course this is not your fault, but the retrace obtained from your report doesn't provide useful information for us as it seems to be corrupted. For this reason, we are now going to close the bug report - please feel free to reopen when you have more information at hand.

And please don't doubt to report any other crashes/issues you might have in the future.

Further information can be found at [1].

[1] https://wiki.ubuntu.com/MozillaTeam/Bugs

Changed in firefox:
status: Incomplete → Invalid
Revision history for this message
Markus Plessing (markus-plessing) wrote :
Changed in eclipse:
status: Unknown → Confirmed
Revision history for this message
In , Ericwill (ericwill) wrote :

I've never seen this crash. Closing, please feel free to re-open it if someone sees it again.

Changed in eclipse:
importance: Unknown → Medium
status: Confirmed → Fix Released
Changed in eclipse:
status: Fix Released → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.