This is a startup crash with Thunderbird 0.9. Here is the latest from Talkback:
Count Offset Real Signature
[ 144 nsXULDocument::OnStreamComplete e20cd02c - nsXULDocument::OnStreamComplete ]
Crash date range: 05-NOV-04 to 15-NOV-04
Min/Max Seconds since last crash: 0 - 132118
Min/Max Runtime: 0 - 134524
Count Platform List
126 Windows XP [Windows NT 5.1 build 2600]
18 Windows 2K [Windows NT 5.0 build 2195]
Count Build Id List
No of Unique Users 128
(1949077) Comments: just wanted to launch it : I upgraded from 8 to 9.
Thnderbird launch and when loaded the bottom third of the screen is loaded with
a kind of debug information writen in red
(1943080) Comments: I was lauching Thunderbird when a dialog box appears.
(1940138) Comments: Ich habe Thunderbird neu installiert - aber kein Start
(1939789) Comments: Thunderbird would not start instead a white window wih
red written indication is shown
(1939240) Comments: starting the newest (german) version...
(1926971) Comments: heu nothing !
(1922846) Comments: I had the thunderbird engl. version installed. when I
downloaded the german version and had it installed it does not work anymore.
(1921835) Comments: Trying to open mail from firefox
(1920526) Comments: only opening the mail application
(1919967) Comments: after installation of update 0.9 the program starts
just 1 time after i have the following display:
-------------------------------^ <menuitem id="menu_close"
(1919967) Comments: accesskey="&closeCmd.accesskey;"
<mail-urlfield keywordrelated="true" id="expandedcontent-baseBox"
<key id="key_groupBySort" key="&groupBySortCmd.key;"
(1912991) Comments: aggiornamento 0.73 - 0.9
(1906863) Comments: no mails are shown anymore. After the update to 0.9.
Istallationfile: Installed the thunderbird-0.9.de-DE.win32.installer.exe
Theres 2 lines of code shown below. The pointer in the first line is at:
label="&closeCmd.label;" the second at:
(1906863) Comments: key="&groupBySortCmd.key;"
(1904788) Comments: starting the application
(1902049) Comments: was just starting it
(1900777) Comments: just started thunderbird
(1892921) Comments: I was just running 'azureus' and trying to open thunderbird
(1885121) Comments: Pressed "Check eMail" via Keyboard
(1882973) Comments: starting Thunderbird
(1881908) Comments: When opening Thunderbird .; Impossible to use !
(1872778) Comments: I installed thunderbird 0.9 over 0.8; if i start via
the prompt of installation procedure everything works; afterwards the attempt
to start again via desktop entry fails; reboot on other system had worked on
this did not
(1850790) Comments: Starting thunderbird
(1848734) Comments: Launch Thunderbird 0.9
(1845977) Comments: The failure occured where I start Thunderbird
(1842093) Comments: Updated to 0.9 from (I think) 0.8 (it was not
deinstalled). Before reboot Thunderbird worked just fine. After reboot: no
menubar no toolbar crazy chars at bottom then this.
(1833094) Comments: i'm trying to open FireBird
(1824880) Comments: i startet it
(1805350) Comments: I've done the update it had work one time with it and
now I can't launch Thunderbird and get my mail Please help me if you can thanks
(1802657) Comments: i just opend FB - there were no statusbar or menue.
and at the bottom there were a failure-msg something like "key... label..."..
(1797162) Comments: I just wanted to start the programm but then an
error-message was shown and Mozilla Thunderbird stopped.
(1782662) Comments: I closed Thunderbird and re-opend it. No mails
postings... was seen. All has gone to hell :-(
(1767949) Comments: installed new localized version it worked. Closed
Thunderbird opened later again - nothing working anymore (thunderbird opened
but i couldnA't do anything my mails couldnA't be seen...). Restarted Computer
but no change
Another crash with a similar stack signature and trace was fixed a while back,
so perhaps it can provide some clues to what's going on here. See bug 204935.
Please update the component if necessary.
Could this have to do with an incompatible extension?
Considering that most user comments mention upgrading from a previous version,
incompatible extensions might be the cause. I'll keep an eye on the Talkback
data to see if any mention of extensions show up.
i suspect there's nothing we can do here for this subset of users.
this is usually the result of folks that were using version of thunderbird
before we had an extension manager system that disabled extensions when
upgrading. So they have extensions from 0.4, 0.5 that aren't versioned at all
that are still getting loaded and bad things happen.
jay: can you contact them? XUL.mfasl might be of interest.
Hmm... just looking at the email addresses of these users, it looks like they're
all from outside the US. I see .de, .ch and .fr addresses. Any locale specific
issues that might be causing this crash?
For those that have access to the internal Talkback query tool (Scott), try this:
I'll try contacting a few of them to see if they're willing to post more info in
could be a bad locale, i suppose. i'd really want to see stuff before i tried to
make a guess.
I used 0.8 version in EN and I instaled the 0.9 version in FR. after instalation
thunderbird dosen't work (crash, bad inerface, ...).
Since I instaled the 0.9 version in EN and this version work good.
I apologize for my english language (I'm french and if somebody in the teams
speaks fluently french, that might help...)
I was using the zip file installation of thunderbird since the 0.3 and this
time (for the late 0.9) I had taken the windows intaller the first launch was ok
but the second time it crashed showing me a strange view of the application
I was using extensions
- Easy get mail button O.2 (I think it already didn't work with the 0.8)
- Minimize to tray 0.0.1.20040828+
I might help you better if you tell me what you're looking for...
I'm guessing this crash has something to do with users installing some localized
version of Thunderbird over an old en-US build in the same location. Is there
anything we can do to prevent things from breaking or at least avoid the crash?
It does seem like a bad idea to do a pave over install from en-US to something
else, but a lot of users are doing it anyway...
This seems to be a topcrasher for Firefox (mostly MacOSX) and Mozilla builds as
well. All comments mention updating to a recent build (with a few extensions
installed). This seems to be a general problem with installing over an older
build with extensions that may or may not be outdated. Updating component to
cover all products.
His a recent set of crashes for the MozillaTrunk:
Count Offset Real Signature
[ 3 nsXULDocument::OnStreamComplete e20cd02c - nsXULDocument::OnStreamComplete ]
Crash date range: 19-JAN-05 to 24-JAN-05
Min/Max Seconds since last crash: 2 - 3714
Min/Max Runtime: 2 - 3714
Count Platform List
3 Windows XP [Windows NT 5.1 build 2600]
Count Build Id List
No of Unique Users 3
And a link to all nsXULDocument::OnStreamComplete crashes across products and
Scott: We should probably pick a more detailed component and find a new owner
for this bug.
Jay, does this crash still appear? I suspect that bug 278534 would at least get
rid of the bad locale registration issues, although we should probably fix the
core underlying problem here as well.
Benjamin: Yes, this crash is happening with Trunk builds as recent as
yesterday's. Let me know when your patch in bug 278534 is approved and checked
in. I can keep an eye on Talkback data after that to see if it helps crash
numbers go down.
So looking at the talkback stack, the only thing I can assume is that it's off
by 2 lines and the script proto is null or bogus.
Jay, is there anything you can do as far as getting me locals here?
Boris: Talkback is often off by 2 lines, so you're probably right about that.
I also took a look at detailed reports from Talkback on our internal server and
something is broken...I didn't see any stacks at all and therefore couldn't find
any local var info or param lists (which normally is missing anyway). Sorry.
I'll look into the detailed report problem though.
Is bug 280157 a duplicate of this bug? Just asking.
*** Bug 280157 has been marked as a duplicate of this bug. ***
This continues to be a topcrasher and after looking into the detailed Talkback
reports problem, I have not figured out a way to get the local var info.
Hopefully someone will be able to figure out what is going on here with the data
that we do have:
I have not been able to reproduce this, but if someone else can and knows what
specific info from users will be valuable, I can setup our feedback system to
contact them and ask for more help.
Marking topcrash+ and nominating for Aviary 1.1 to get some traction.
Reassigning to default owners since this isn't really just a Thunderbird crash.
nobody isn't much of an improvement over mscott ...
still crashing in Mozilla 1.6b1 on WinXP sp1, here the talkback ID:4840306 and
today my thunderbird (nightly build 2006010205) crashes every time I try to start it.
The funny thing is that if I start it with "-profilemanager" no profiles show up in the list.
Still frequent talkback - the interesting FF ones below (no thunderbird stacks are interesting)
TB26804437 FF2 linux
TB28442911 FF2 win restarting after installing skype beta extension...
TB28648803 trunk mac
Stack Signature nsXULDocument::InsertElement() 5c9642ec
Product ID FirefoxTrunk
Build ID 2006120408
Trigger Time 2007-01-23 19:16:24.0
Operating System Darwin 8.8.0
Module firefox-bin + (001abc00)
Since Last Crash 345 sec
Total Uptime 345 sec
Trigger Reason SIGBUS: Bus Error: (signal 10)
Source File, Line No. /builds/tinderbox/Fx-Mozilla1.9-Release/Darwin_8.7.2_Depend/build/universal/ppc/db/sqlite3/src//builds/tinderbox/Fx-Mozilla1.9-Release/Darwin_8.7.2_Depend/mozilla/db/sqlite3/src/pragma.c, line 846
nsXULDocument::InsertElement() [mozilla/db/sqlite3/src/pragma.c, line 846]
nsXULDocument::InsertElement() [mozilla/db/sqlite3/src/pragma.c, line 4289]
nsXULDocument::OverlayForwardReference::Resolve() [mozilla/db/sqlite3/src/pragma.c, line 846]
nsXULDocument::ResolveForwardReferences() [mozilla/db/sqlite3/src/pragma.c, line 1064]
nsXULDocument::ResumeWalk() [mozilla/db/sqlite3/src/pragma.c, line 3069]
nsXULDocument::OnStreamComplete() [mozilla/db/sqlite3/src/pragma.c, line 3460]
nsStreamLoader::OnStopRequest() [mozilla/netwerk/base/src/nsStreamLoader.cpp, line 713]
nsJARChannel::OnStopRequest() [mozilla/modules/libjar/nsJARChannel.cpp, line 713]
nsInputStreamPump::OnStateStop() [mozilla/db/sqlite3/src/pragma.c, line 713]
nsInputStreamPump::OnInputStreamReady() [mozilla/db/sqlite3/src/pragma.c, line 396]
nsThread::ProcessNextEvent() [mozilla/xpcom/threads/nsThread.cpp, line 483]
NS_ProcessPendingEvents_P() nsBaseAppShell::Run() [mozilla/widget/src/xpwidgets/nsBaseAppShell.cpp, line 156]
-[AppShellDelegate runAppShell]() [/builds/tinderbox/Fx-Mozilla1.9-Release/Darwin_8.7.2_Depend/build/universal/ppc/db/sqlite3/src//builds/tinderbox/Fx-Mozilla1.9-Release/Darwin_8.7.2_Depend/mozilla/db/sqlite3/src/pragma.c, line 419]
- nsAppShell::ProcessNextNativeEvent() [mozilla/db/sqlite3/src/pragma.c, line 273]
nsBaseAppShell::OnProcessNextEvent() [mozilla/widget/src/xpwidgets/nsBaseAppShell.cpp, line 2]
nsThread::ProcessNextEvent() [mozilla/xpcom/threads/nsThread.cpp, line 496]
NS_ProcessPendingEvents_P() nsBaseAppShell::NativeEventCallback() [mozilla/widget/src/xpwidgets/nsBaseAppShell.cpp, line ]
nsAppShell::ProcessGeckoEvents() [mozilla/db/sqlite3/src/pragma.c, line 198]
-[AppShellDelegate handlePortMessage:]() [/builds/tinderbox/Fx-Mozilla1.9-Release/Darwin_8.7.2_Depend/build/universal/ppc/db/sqlite3/src//builds/tinderbox/Fx-Mozilla1.9-Release/Darwin_8.7.2_Depend/mozilla/db/sqlite3/src/pragma.c, line 409]
- - nsAppShell::Run() [mozilla/db/sqlite3/src/pragma.c, line 316]
nsAppStartup::Run() [mozilla/db/sqlite3/src/pragma.c, line 172]
XRE_main() [mozilla/toolkit/xre/nsAppRunner.cpp, line 2515]
*** Bug 306910 has been marked as a duplicate of this bug. ***
No topcrasher anymore but still happens with trunk builds:
seems to be gone. not reports for:
- Thunderbird trunk (last 2 weeks) and TB2.0.18, 0 crashes on talkback
- FF last 8 days .... 3.0.1-4, 3.0b1-5, last 12 days 3.1b2 and pre
I don't think, it's gone. This issue occured to me out of sudden after upgrading from Fx 3.0.4 to Fx 3.0.5. Crash Report: http://crash-stats.mozilla.com/report/index/8cceeeb5-2b0d-45ea-a8ad-093882081226?p=1
That's true. It's back in the topcrasher list as #9 for Firefox 3.0.5. There are about 1550 reported crashes mainly on Windows but also OS X. Even for Firefox 3.1 beta 2 it's listed as #45 with 140 crashes but only on Windows.
Tobias, could you install the 3.0.4 version and try to upgrade again? It would be interesting to see, if it's reproducible.
> Tobias, could you install the 3.0.4 version and try to upgrade again? It would
> be interesting to see, if it's reproducible.
This issue did only occur on my father's PC (Win XP). I'm now back home and I won't have access to it this year, anymore. Maybe someone else can help?
(In reply to comment #28)
> This issue did only occur on my father's PC (Win XP). I'm now back home and I
> won't have access to it this year, anymore. Maybe someone else can help?
I've no idea if this is profile related. You could ask your father to let him send his profile to you? So you could check it out? That would be really helpful.
(In reply to comment #29)
> I've no idea if this is profile related. You could ask your father to let him
> send his profile to you? So you could check it out? That would be really
Okay, I now fetched his profile folder and installed Fx. 3.0.4 in a different folder on my system, then imported the profile and updated to 3.0.5. What I have to tell is that it does NOT crash on the Win Vista computer, the update completed successfully - no crash.
If it's helpful, I could attach his profile folder to this bug. But only if it's deleted as soon as you don't need it anymore.
(In reply to comment #30)
> If it's helpful, I could attach his profile folder to this bug. But only if
> it's deleted as soon as you don't need it anymore.
Please don't do that! Attachments cannot be deleted anymore.
I've taken a look at the crash report again and noticed that in frame 2 we are using a nsJARChannel. Do we probably crash while accessing a jar file, e.g trying to fetch the version information of extensions? The following crash report could be a bit more helpful: bp-3bfe1499-e1ed-47f7-bd40-71c9d2081229
This is #18 of top crasher bugs for Fx3.1b2 and #9 for Fx3.0.5 now. Looks like more and more users are affected by this bug.
Tobias, which extensions are your father using?
(In reply to comment #31)
> Tobias, which extensions are your father using?
As far as I can tell from the profile, he's using the following extensions:
IE Tab 1.5.20081203
MediaPlayer Connectivity 0.9.1
Yahoo! Toolbar 184.108.40.20680910
Fasterfox 2.0.0 (made compatible by myself increasing the em:MaxVersion value)
Adblock Plus 1.0
Deutsches Wörterbuch 1.0.1
So it was gone for a few days and then reappeared when we released 3.0.5? So does it happen once when certain people upgrade, perhaps?
Maybe something bad gets in the profile, then they upgrade and crash, then the profile data is fixed at least until the next upgrade?
I hate this bad old bug but I have no idea how to make progress.
*** Bug 473174 has been marked as a duplicate of this bug. ***
I have no idea whether this is helpful, but since everybody seems to be at a bit of a loss, I'll just throw it in:
Looking at one of the crash reports associated with my original bugreport (Bug 473174), I noticed that it says: Crash Reason EXCEPTION_ACCESS_VIOLATION. Now, I may be misinterpreting this of course, but if it means that there was an access violation trying to access some memory location, it may be related to some other issue that has been getting on my nerves occasionally. Among other plugins, I have installed Adobe's Acrobat plugin in what I believe is a pretty recent version. (Not sure how to find out which version it is, about:plugins doesn't seem to show the version number). Sometimes, when I have a PDF open in the browser and then close the browser, I'll get a Windows message box saying that there was an access violation accessing address "0x..." (sorry, haven't got the exact address at hand) and that the memory location could not be "read". Other than that, Firefox will shut down quite normally, though.
As I said, no idea if this is related.
steffen: are you on Windows? If so, are you willing to use a debugger to help track down the cause? Nobody is working on this bug because no developer can reproduce it. If you'd like to try with a debugger, there are docs here:
If you join irc.mozilla.org #developers, someone may be willing to help you through the process.
Steffen, further please test with all add-ons disabled or better in Safe Mode. Does it happen also in this mode? If yes, does a new profile help?
Sorry, things have been a bit busy at my side, so I haven't actually been able to respond. Yes, I am on Windows. I was willing to do the debugger thing, but as I said, I got swamped. However, strangely enough, for some time now, I have been able to start Firefox without problems. I am currently running 3.0.7 and fearing to update to 3.0.8 lest I get the same problems again.
However, having said that, all is still not well, as some 2 out of 3 times I start Firefox, it will not start correctly some of the plugins I have installed. In particular, I am having problems with NoScript (currently version 220.127.116.11) and TabMixPlus (currently at version 0.3.7pre.080728). The behaviour I observe is as follows:
- For NoScript, the NoScript icon does not appear in the lower-right corner
potentially dangerous contents is still blocked. I just don't get a chance
to unblock it if I need to.
- For TabMixPlus, the easiest way to spot the bug is by looking for the
close-tab button. I have configured TabMixPlus to place these on the tab
itself, so when there is only one button at the very far right of the tab
row, I know something has gone wrong. In this case, I will not be able to
close any newly created tabs, even using the button that is in the wrong
Restarting Firefox often resolves the problem. However, note that the two failures seem not to be coupled. It is possible to have one of the plugins working correctly but not the other one.
Sorry again for the delay in responding,
Steffen, this bug is about the crash and not for any issues with both add-ons. Please talk with the add-on authors for those problems. If you have helpful information regarding the crash please report back. If it will happen again please follow my comment 38. Thanks.
Thanks for the reply. I am aware that this is a bug about the crash, but maybe wasn't sufficiently clear about my reasons for talking about the plugins here.
I mentioned the plugin issue as it seems to be co-located with this bug (see also my comment 36 above). So, I was hoping that this behaviour may give Firefox developers some more insight into this bug. It looks like a race condition to me, and as we all know these can be fiendishly difficult to debug. At least in my experience, any information about weird behaviour that seems to be co-located may be helpful in figuring out the true source of the bug.
Perhaps bug 497875 is related to this bug?
*** Bug 475025 has been marked as a duplicate of this bug. ***
this is going to want some attention in the next couple weeks to determine if this is TB3 needs. 3.0b3 topcrash if the TB crash stats are stable in the next several days.
FF current trunk line# differs by one vs TB based 1.9.1
- FF trunk bp-e027e6c4-eb2e-4089-ac38-c98bb2090720
- TB 3.x bp-5205ccdf-4961-4c14-8b95-64f712090722
Created attachment 392930 [details] [diff] [review]
(In reply to comment #44)
> this is going to want some attention in the next couple weeks to determine if
> this is TB3 needs. 3.0b3 topcrash if the TB crash stats are stable in the next
> several days.
not even in the top 100 for 3.0b3 or 3.0b4pre unless I misread the stats, so not a great concern for thunderbird
Just go this crash today launching Minefield (3 consecutive crashes):
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.3a1pre) Gecko/20090903 Minefield/3.7a1pre ID:20090903045002
I wonder if my crash could have been triggered by bug 514131
Frame Module Signature [Expand] Source
0 xul.dll nsXULDocument::OnStreamComplete content/xul/document/src/nsXULDocument.cpp:3523
1 xul.dll nsStreamLoader::OnStopRequest netwerk/base/src/nsStreamLoader.cpp:108
2 xul.dll nsJARChannel::OnStopRequest modules/libjar/nsJARChannel.cpp:880
3 xul.dll xul.dll@0x98cc6f
4 xul.dll nsInputStreamPump::OnInputStreamReady netwerk/base/src/nsInputStreamPump.cpp:401
5 xul.dll nsOutputStreamReadyEvent::Run xpcom/io/nsStreamUtils.cpp:191
6 xul.dll nsThread::ProcessNextEvent xpcom/threads/nsThread.cpp:527
7 xul.dll nsBaseAppShell::Run widget/src/xpwidgets/nsBaseAppShell.cpp:170
8 xul.dll nsAppStartup::Run toolkit/components/startup/src/nsAppStartup.cpp:182
9 nspr4.dll PR_GetEnv
10 firefox.exe wmain toolkit/xre/nsWindowsWMain.cpp:110
11 firefox.exe firefox.exe@0x21a7
12 kernel32.dll BaseProcessStart
FWIW, I got this sequence of crashes updating from yesterday's minefield win32 zip to today's:
and trying to upgrade to firebug 1.5x.0a23 (from 1.5x.0a22).
The recent cluster of these seems to be confined to the 20090903 nightly build. I suspect the patch for bug 510844 somehow aggravated this issue.
Note, for the 5 crashreports mentioned in #48, only one is related to nsJARInputStream, so it is doubtful that bug 510844 is the cause of this.
On 20090903 there a number of other patches committed and backed out because of tinderbox issues.
#98 crash for Thunderbird v3.0.0, so if reporters confirm this is fixed with trunk build and the patch isn't offensive, suggest this is wanted for 3.0.1.
Comment on attachment 392930 [details] [diff] [review]
(guess it can't make 18.104.22.168=TB 3.0.1)
Comment on attachment 392930 [details] [diff] [review]
Approved for 22.214.171.124, a=dveditz for release-drivers
This appears for me in Fx 3.6, but only if NoScript is enabled. When will the wallpaper be brought to the 1.9.2 branch?
Comment on attachment 392930 [details] [diff] [review]
Approved for 126.96.36.199, a=dveditz for release-drivers
What is the scenario to reproduce this bug pre-fix? I see that No Script is involved in once instance but no details.
QA has nothing to do here to verify this fix.
I can confirm that my problem is fixed in 3.6.2.