528153 - karmic im-uim crashes firefox [@ libgobject-2.0.so.0.2200.2@0x2909a - im-uim.so@0x5f01]

Status: VERIFIED INVALID

(Firefox :: General, defect)

Description

[Dave M G]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.5) Gecko/20091109 Ubuntu/9.10 (karmic) Firefox/3.5.5
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.5) Gecko/20091109 Ubuntu/9.10 (karmic) Firefox/3.5.5

On the advice of someone on Launchpad (The Ubuntu bug reporting site), I am cross posting this bug here.
https://bugs.launchpad.net/ubuntu/+source/firefox-3.5/+bug/477513

FireFox crashes routinely. 

Reproducible: Always

Steps to Reproduce:
1. Open FireFox
2. Surf to various sites. That's all that's needed.

Actual Results:  
FireFox will crash.

Expected Results:  
FireFox should not crash.

Seems to be related to changing pages. If I stay on a site and don't navigate to too many pages, FireFox may stay alive for longer (though not guarunteed).

Surfing to many pages in a shorter period of time increases the likelihood of crashing sooner.

I can reliably crash FireFox in under a minute by randomly surfing without stopping to look too long at the content.

Please see the bug report on launchpad for crash output and further details.

Comment 1

[Dave M G]

For some reason the Launchpad people made me change bug reports.

The new URL is here:

https://bugs.launchpad.net/ubuntu/+source/firefox-3.5/+bug/481237

Comment 2

[Tyler Downer [He/Him]]

You seem to be using a ubuntu build of firefox. These are known to occasionally have their own bugs. Please try using a build from mozilla.com. Once you get one, if it still crashes, you can get a stacktrace, https://developer.mozilla.org/En/How_to_get_a_stacktrace_for_a_bug_report

Comment 3

[Matthias Versen [:Matti]]

THe ubuntu build should be nearly identical to Mozila.org build. The crash reporter is only working in Mozilla.org builds, either use such a build or use the Ubuntu instructions in that link to provide a stack trace.

Comment 4

[Dave M G]

I tried downloading a build directly from the FireFox web site and running that.

It went for a while without crashing and I thought maybe it would hang tough.

But then it crashed as well. A bug report thing came up, but I'm not sure if it went to this site or not though.

So, anyway, looks like the bug might be FireFox related, not necessarily to do with the Ubuntu build.

Comment 5

[timeless]

https://developer.mozilla.org/En/How_to_get_a_stacktrace_for_a_bug_report

Comment 6

[Dave M G]

I sent two stacktrace reports from FireFox. But now that they are uploaded to this web site, I can't find them to link to them here.

So please don't just tell me I need to file this or that report. I'm trying to give you what you need to fix the bug.

Comment 7

[Tyler Downer [He/Him]]

Even though you have submitted the stacktraces, we get thousands of them every day. So, we have no idea which is yours. If you read the link it give instructions on how to go and grab the id so you can past it here.

Comment 8

[Dave M G]

Yeah, so the instructions say:
"To access all of your submitted reports type "about:crashes" into the Firefox address bar and press enter"

So I type "about:crashes", without the quotes, into the Firefox address bar, and press enter, and it says:

"The URL is not valid and cannot be loaded."

So what now?

Comment 9

[Dave M G]

Alright, I got the crash report IDs. I don't know what changed, but after crashing Firefox a couple more times, the "about:crashes" thing magically started working.

bp-b43b361f-87cb-411a-94ee-fb3432091116	09-11-16	11:49 PM
bp-716ba922-b40c-4770-b236-1dec12091116	09-11-16	11:48 PM
bp-b86b95d0-d8ee-4a95-a7d9-0e6e62091116	09-11-16	11:34 PM
bp-cce946ca-cd63-4e8a-b52b-6c36e2091115	09-11-16	12:16 AM
bp-db3ba632-519b-4222-a3d1-fa57b2091114	09-11-15	04:32 AM

Comment 10

[timeless]

Signature	libgobject-2.0.so.0.2200.2@0x2909a
UUID	b43b361f-87cb-411a-94ee-fb3432091116
Time 	2009-11-16 06:49:15.93088
Uptime	23
Last Crash	65 seconds before submission
Product	Firefox
Version	3.5.5
Build ID	20091102141836
Branch	1.9.1
OS	Linux
OS Version	0.0.0 Linux 2.6.31-15-generic #50-Ubuntu SMP Tue Nov 10 14:54:29 UTC 2009 i686 GNU/Linux
CPU	x86
CPU Info	GenuineIntel family 13 model 23 stepping 10
Crash Reason	SIGSEGV
Crash Address	0x9c209a
User Comments	
Processor Notes 	
Related Bugs

Crashing Thread
Frame 	Module 	Signature [Expand] 	Source
0 	libgobject-2.0.so.0.2200.2 	libgobject-2.0.so.0.2200.2@0x2909a 	
1 	im-uim.so 	im-uim.so@0x5f01 	
2 	im-uim.so 	im-uim.so@0x6b34 	
3 	im-uim.so 	im-uim.so@0x6c10 	
4 	libgtk-x11-2.0.so.0.1800.3 	libgtk-x11-2.0.so.0.1800.3@0x11c403 	
5 	libgtk-x11-2.0.so.0.1800.3 	libgtk-x11-2.0.so.0.1800.3@0x11c403 	
6 	libxul.so 	nsWindow::IMESetFocus 	widget/src/gtk2/nsWindow.cpp:6418
7 	libxul.so 	nsWindow::SetFocus 	widget/src/gtk2/nsWindow.cpp:1454
8 	libxul.so 	nsGlobalWindow::Focus 	dom/src/base/nsGlobalWindow.cpp:4314

im-uim.so 		8DF4E7F64B5A5088C8F022A809F835C80 	im-uim.so

So, you have an input method of sorts. And it's crashing. I'm sorry that you have this software installed. Please file a bug against it.

To debug this further you will have to install symbols for:
libgobject-2.0.so.0.2200.2
im-uim.so
libgtk-x11-2.0.so.0.1800.3

(these should come from your "karmic" vendor)
and then you should take the generated crash report (good luck) and report it in your "karmic" vendor's bug reporting system.

Alternatively, I'd suggest you find a way to disable im-uim, as I'm certain that the vast majority of our users aren't using it.

Comment 11

[Dave M G]

I'm not sure I understand your claim that it is UIM that is crashing.

*Only* Firefox crashes. I use UIM/Anthy to switch input methods when I type in Japanese. I use UIM/Anthy in every software package I have where I type in text, and Firefox is the only software that crashes on my system.

Further, not once has Firefox crashed when I have been inputting Japanese. It has only ever happened when I open and close pages.

In fact, for the times that I have been testing the Firefox build that you recommended I download straight from you instead of from the Ubuntu repositories, I have never even typed anything in, at all, ever. I would open it, surf to random web sites until it crashed, and that's it.

If your claim is that the mere presence of UIM on my system is enough to cause Firefox to crash, regardless of whether or not I am actually using it, then I suppose that might be the case, as that would be consistent with my experience.

However, since UIM has no other problem with any other software, I think that puts the onus on Firefox to show why it has a problem, not UIM.

Comment 12

[Matthias Versen [:Matti]]

A stacktrace shows the last steps of an application before the application crashed and the first line in the stack trace is the position that crashed.
The stacktrace clearly shows that Firefox isn't the cause of the crash, it's either bug in libgobject-2.0.so.0.2200.2 (frame0) or more likely im-uim.so (frame1-3)

The last Firefox code is in frame6 and that means that Firefox code can not be the cause of the crash. Every Developer who knows C, C++ will confirm this.

BTW: I'm not a developer but I learned to parse Stack traces to find the right bugzilla products/components.

Comment 13

[timeless]

I'm not sure why you're coming here asking for help (which is not our job, we are not http://support.mozilla.com) without being willing to accept our analysis.

If you'd like to learn about how debuggers work and how to read stack traces, I'll gladly give you lessons, however if you're going to act like an end user, then please do us the favor of trusting us.

It's rude to ask a doctor "please diagnose my problem" and when he says "you have the flu", say "oh, why should i believe you?"

Note that I've already provided instructions you can follow to analyze this for yourself. I'll repeat them because I'm feeling generous.

> To debug this further you will have to install symbols for:
> libgobject-2.0.so.0.2200.2
> im-uim.so
> libgtk-x11-2.0.so.0.1800.3

> (these should come from your "karmic" vendor)
> and then you should take the generated crash report (good luck) and report it
> in your "karmic" vendor's bug reporting system.

Please keep in mind, we're under no obligation, please read:
  https://bugzilla.mozilla.org/page.cgi?id=etiquette.html

If you're willing to trust our documentation, you're free to read:
https://developer.mozilla.org/En/Debugging_Mozilla_on_Linux_FAQ

It's the google top hit for "Linux Debugging FAQ" and "Unix debugging faq" so perhaps it's trustworthy. OTOH, you're free not to trust our content.

Resolved: INVALID. Do not comment without a stack trace that includes debugging symbols for the crashing frames. Most likely you will be best able to achieve this by asking your vendor for support. However, this bug has more than enough information for you to do it yourself.

Comment 14

[Dave M G]

You make a fair point and I'll tell you exactly why I am arguing your analysis.

I am a programmer and I do debugging as well, and I am not unfamiliar with the complexities of debugging. I know full well that just because a problem manifests down the line does not mean the problem was further up the code.

I am also fully aware that in situations where people are doing technical support for a lot of people, their first objective is to keep their inbox clear by finding someone else to blame.

You raise the analogy of a doctor. If I went to a doctor and I said my foot hurt, and he did some kind of scan of my body and said the problem was my hand, I would argue the point even though I am not a doctor.

You want me to go to the people who code UIM and say "I need to report a bug with your software that is not activated by actually interacting with your software in any way, does not manifest in any other software but Firefox, and only causes Firefox to crash when I'm using Firefox without touching any of your components".

I wouldn't blame them if they said "you know, I think your problem might be Firefox..."

Comment 15

[Matthias Versen [:Matti]]

We don't do support here, bugzilla is a place for the developers.
The difference between support and us is that the support tries to help if people have issues with software and it doesn't matter what caused the problem in the first place. In bugzilla we try to identify bugs in the code of our product and fix the bugs in the code.

Your crash is identified by 2 independent people as bug in im-uim.so and not in our code. Sorry if you are not happy with he resolution of this bug (invalid) but do not change it back. 
- https://bugzilla.mozilla.org/page.cgi?id=etiquette.html -> point 2.2

>You want me to go to the people who code UIM and say "I need to report a bug
>with your software that is not activated by actually interacting with your
>software in any way, does not manifest in any other software but Firefox, and
>only causes Firefox to crash when I'm using Firefox without touching any of
>your components".

Sure that is expected by us, and you should post them the stack trace as argument and they should confirm the case.

And finally you can thrust timeless words more than my words. He is a very respected developer earns his money with developing software (not paid by Mozilla.org) and he analyzed and fixed hundreds of Gecko crashes in the past while I'm only a bug triager in bugzila for several (8+) years .

Comment 16

[Reed Loden [:reed]]

Take bug 418073 for example... Firefox crashed, but the cause of the crash was clearly im-uim, so a fix upstream at the im-uim level was required.

Comment 17

[timeless]

http://packages.ubuntu.com/karmic/uim-gtk2.0 indicates the version of ui-uim should be 1.5.6

Comment 18

[Dave M G]

I have reported the bug to the UIM developers:

http://bugs.freedesktop.org/show_bug.cgi?id=25139

Comment 19

[Etsushi Kato]

I've already fixed the bug in im-uim.so, and uim 1.5.7 will be released this weekend.  Thanks for notifying me of the bug.

Comment 20

[Dave M G]

Fix confirmed!!

I was able to download the new fix through a Debian package offered by an Ubuntu developer's PPA, and I have been torture testing Firefox for the last hour or so. There was one freeze (probably because of the web site), but no crashes.

I ran Firefox from the command line, and there was no output whatsoever indicating any errors, which is completely different than what I was experiencing before.

I am confident enough in Firefox's performance now to say that for me the bug seems to be fixed with the updated UIM, and Firefox is no longer unstable.

Thank you to everyone who helped to diagnose and fix the problem. Your hard work is much appreciated.