Closed Bug 421929 Opened 16 years ago Closed 14 years ago

firefox crash on Google Reader [@ js_GetGCStringRuntime]

Categories

(Core :: JavaScript Engine, defect)

defect
Not set
critical

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: mcepl, Unassigned)

References

()

Details

(Keywords: crash)

Crash Data

Attachments

(1 file)

User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9b5pre) Gecko/2008030918 Fedora/3.0b5pre-0.37.cvs20080309.fc9 Minefield/3.0b5pre
Build Identifier: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9b5pre) Gecko/2008030918 Fedora/3.0b5pre-0.37.cvs20080309.fc9 Minefield/3.0b5pre

Occasionally going from a website to gmail or google reader firefox will crash
with the bugbuddy dialog.  After restarting firefox, it seems to work fine for
several hours.

I thought it might be flash, but in most cases going to gmail or google reader
doesn't require flash, but those application use quite a bit of Javascript,
so is it possible for Javascript to crash firefox?

Version-Release number of selected component (if applicable):
[jesusr@roadrunner ~]$ rpm -q firefox
firefox-2.0.0.12-1.fc8.i386


How reproducible:
At least once or more times daily.

Steps to Reproduce:
1. Open firefox
2. Browse to a website www.redhat.com, zeusville.wordpress.com, or gmail.
3. After perusing the site for a bit, go to http://www.google.com/reader/view/
4. Before the page loads firefox crashes with bugbuddy. 
5. Restarting firefox and choosing restore session will NOT restore the
   url from #3 as it doesn't think it actually loaded it.
  
Actual results:
firefox goes BOOM

Expected results:
firefox continues to behave and be my slave app handling countless hours
of abuse :) or in other words, it should just work.



Reproducible: Sometimes

Steps to Reproduce:
Here is the gdb output. scenario used to create crash:
1) start firefox (displays start.fedoraproject.org)
2) click on Gmail bookmark (http://mail.google.com)
3) while gmail is loading I double click to create a new tab
4) once gmail is loaded, I click the google reader bookmark
(http://www.google.com/reader/view/)
5) firefox segv :)
Actual Results:  
crash

Expected Results:  
shouldn't

Hi, this was originally filed into Red Hat bugzilla. I have tried to reproduced this, and firefox-3.0-0.37.cvs20080309.fc9.x86_64 really crashed while importing OPML file. However, after I have downloaded all -debuginfo files and restarted firefox, it hasn't crash for me.
Thread 1 (Thread -1208465712 (LWP 12127)):
#0  0x00b6760c in js_GetGCStringRuntime (str=0x0) at jsgc.c:508
js_GetGCStringRuntime(0);
js_GetStringBytes(0, 0);
JS_GetStringBytes(0);
JS_GetFunctionName(JSFunction *fun) fun->atom && !ATOM_TO_STRING(fun->atom)
shaver suggests fun->atom == STRING_TO_ATOM(NULL)
#1  0x00b39e30 in JS_GetFunctionName (fun=0xa5a9f10) at jsapi.c:3586
#2  0x0063f046 in jsd_GetScriptFunctionName (jsdc=0x95eb1c0, jsdscript=0xb4e9f40)
    at jsd_scpt.c:462
#3  0x0063c558 in JSD_GetScriptFunctionName (jsdc=0x95eb1c0, jsdscript=0xb4e9f40)
    at jsdebug.c:291
#4  0x00648874 in jsdScript (this=0x0, aCx=0x95eb1c0, aScript=0xb4e9f40)
    at jsd_xpc.cpp:984
#5  0x00649286 in jsdService::EnumerateScripts (this=0x95eb178, 
    enumerator=0xae5ae50) at jsd_xpc.h:155
#6  0x06fa3829 in XPTC_InvokeByIndex () at xptcinvoke_gcc_x86_unix.cpp:50
Assignee: nobody → general
Severity: normal → critical
Component: General → JavaScript Engine
Keywords: crash
Product: Firefox → Core
QA Contact: general → general
Summary: firefox crash on Google Reader → firefox crash on Google Reader [@ js_GetGCStringRuntime]
Version: unspecified → Other Branch
Version: Other Branch → 1.8 Branch
(In reply to comment #2)
> Thread 1 (Thread -1208465712 (LWP 12127)):
> #0  0x00b6760c in js_GetGCStringRuntime (str=0x0) at jsgc.c:508
> js_GetGCStringRuntime(0);
> js_GetStringBytes(0, 0);
> JS_GetStringBytes(0);

What do these mean here?

A little more prose, please!
the optimizer omitted frames, I'm merely connecting the dots.
http://crash-stats.mozilla.com/report/index/f4140e78-f948-11dc-89c4-001a4bd43ef6 is a similar stack on trunk, with Firebug 1.1.0b10.  I'd be excited to get that in a debugger, I have to say.
Status: UNCONFIRMED → NEW
Ever confirmed: true
OS: Linux → All
Hardware: PC → All
Version: 1.8 Branch → Trunk
I'm experiencing crashes on Google Reader.  However, the crash only occurs when viewing one of my feeds.  All other feeds work fine.

1.  Add http://feeds.pheedo.com/toms_hardware to your Google Reader RSS Feeds
2.  Click the feed name on Google Reader's left panel
3.  Crash.
I have the same problem with Google Reader.  When I try to open CNBC's Big Idea Blog feed in Goodle Reader, FF freezes.

RSS Feed: http://www.cnbc.com/id/17740822/device/rss/rss.xml

The website opens up just fine on it's own, although it doesn't appear to finish loading.  It appears to be loading something from oimg.nbcuni.com... but never finishes.

I don't get a debugger, or error screen, just a complete hang.

FF3RC2, Vista SP1
ericsprojects@gmail.com: err, no. a hang is not the same as a crash bug. i can't imagine how you managed to pick this bug.

file your own bug, but use:
http://developer.mozilla.org/en/docs/How_to_get_a_stacktrace_with_WinDbg
!analyze -v -hang

to get a stack trace.
I've experienced different symptoms of the problem, all reproduced using the steps mentioned above.  In the early beta's, Firefox simply crashed.  Later, it hung filling up system memory.  Now, the browser hangs consuming CPU.  

I happened to notice a Javascript error about infinite recursion within Firebug.  Unfortunately, I no longer have those details. However, I have "reproduced" the problem on six different machines with various OSs and hardware.

If I have time, I'll try to post a stack trace and those error details. 
Ben indicates his problem has been resolved.
Matej, can you still reproduce this crash?
Whiteboard: [closeme 2010-10-25]
(In reply to comment #11)
> Matej, can you still reproduce this crash?

No, I don't have Google Reader account anymore (nor GMail, and I don't want to create it just for this bug). However, I dimly remember that it was working soon after this.
(In reply to comment #12)
>  I dimly remember that it was working soon after this.

that's sufficient to close your bug WFM.
Status: NEW → RESOLVED
Closed: 14 years ago
Resolution: --- → WORKSFORME
Whiteboard: [closeme 2010-10-25]
Crash Signature: [@ js_GetGCStringRuntime]
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: