Closed Bug 1381945 Opened 4 years ago Closed 4 years ago

Firefox 56.0a1 Crash Report [@ shutdownhang | RtlAnsiStringToUnicodeString | MsgWaitForMultipleObjectsEx | CCliModalLoop::BlockFn ]

Categories

(Core :: General, defect)

Product:
Core
Component:
General
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED INCOMPLETE

People

(Reporter: codacodercodacoder, Unassigned)

Details

Crash Data

Attachments

(2 files)

mozregr.png
95.14 KB, image/png
Details
Mozregression for windows.txt
783 bytes, text/plain
Details 
User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:54.0) Gecko/20100101 Firefox/54.0
Build ID: 20170628075643

Steps to reproduce:

Tear off a tab (window)


Actual results:

See crash reports

https://crash-stats.mozilla.com/report/index/45b5bae8-78dd-4715-ab53-5e6bf0170718

https://crash-stats.mozilla.com/report/index/7a1fb054-074c-468b-ad66-5c4de0170718

100% repeatable.
Adding crash signature to the crash signature field. Looks like there are 6 crashes on nightly/2 installs in the last seven days.

Does the crash happen with any site loaded in the tab?
Status: UNCONFIRMED → NEW
Crash Signature: [@ shutdownhang | RtlAnsiStringToUnicodeString | MsgWaitForMultipleObjectsEx | CCliModalLoop::BlockFn]
Ever confirmed: true
I've tested a few, it's consistent AFAICT.

Annoyingly, it's unable to recover the tabs after killing it and restarting.  It offers the list but says it can't(?)
And btw, it's 56 branch (seems I'm not allowed to change it - which is silly).
FYI, the bug is no longer present.  I'm guessing https://bugzilla.mozilla.org/show_bug.cgi?id=1382043 may have had an effect?
This bug is back - and worse than before...

1 - the browser is now totally hung after beginning the drag operation to tear off the tab

2 - the browser does not detect the crash.


56.0a1 (2017-07-28) (64-bit)
Version: 54 Branch → 56 Branch
I tried to find a regression pointer with mozregression-gui but it crashes out with an error asking me to file a bug. Oh the irony ;)
Update: Bug not present in latest nightly (57 branch) received today.
Hi there. If this comes back, two things to help narrow it down:

- Does this happen if you use the same build, but with a new profile?

- Can you paste the about:support info from a build+profile that reproduces the bug?
Both crash reports contain DropboxExt64.17.0.dll so the culprit is Dropbox.
Component: Untriaged → Other
Product: Firefox → External Software Affecting Firefox
Version: 56 Branch → unspecified
I don't have any extensions - if there's a dropbox addon applied, I did not apply it and don't see an sign of it in the Extensions/Plugins panels.(In reply to :Felipe Gomes (needinfo me!) from comment #8)
> Hi there. If this comes back, two things to help narrow it down:
> 
> - Does this happen if you use the same build, but with a new profile?

It's a clean profile I keep purely for dev work - no plugins/extns

> 
> - Can you paste the about:support info from a build+profile that reproduces
> the bug?

It's my regular dev browser so now that the bug has gone, how would I do that? (note: mozregression crashes on me - see above)
(In reply to Codacoder from comment #6)
> I tried to find a regression pointer with mozregression-gui but it crashes
> out with an error asking me to file a bug. Oh the irony ;)

Can you paste the error here?

Can you try mozregression without the GUI?

(In reply to Kohei Yoshino [:kohei] from comment #9)
> Both crash reports contain DropboxExt64.17.0.dll so the culprit is Dropbox.

Dropbox is there only in the two crash reports from Codacoder, but not in other crash reports with the same signature and from other users. So I think it's probably unrelated.
Component: Other → Untriaged
Product: External Software Affecting Firefox → Core
(In reply to Marco Castelluccio [:marco] from comment #11)
> (In reply to Codacoder from comment #6)
> > I tried to find a regression pointer with mozregression-gui but it crashes
> > out with an error asking me to file a bug. Oh the irony ;)
> 
> Can you paste the error here?

platform: Windows-7-6.1.7601-SP1
python: 2.7.11 FROZEN (32bit)
mozregui: 0.9.2
mozregression: 2.3.2
message: IndexError: list index out of range
traceback:   File ".\mozregui\bisection.py", line 82, in check_merge
  File "..\mozregression\bisector.py", line 241, in handle_merge
  File "..\mozregression\build_range.py", line 94, in __getitem__

I tried a few different start/end points, they all do this (above) eventually.

> Can you try mozregression without the GUI?

Sure.  Link for windows D/L?
(In reply to Codacoder from comment #12)
> (In reply to Marco Castelluccio [:marco] from comment #11)
> > (In reply to Codacoder from comment #6)
> > > I tried to find a regression pointer with mozregression-gui but it crashes
> > > out with an error asking me to file a bug. Oh the irony ;)
> > 
> > Can you paste the error here?
> 
> platform: Windows-7-6.1.7601-SP1
> python: 2.7.11 FROZEN (32bit)
> mozregui: 0.9.2
> mozregression: 2.3.2
> message: IndexError: list index out of range
> traceback:   File ".\mozregui\bisection.py", line 82, in check_merge
>   File "..\mozregression\bisector.py", line 241, in handle_merge
>   File "..\mozregression\build_range.py", line 94, in __getitem__
> 
> I tried a few different start/end points, they all do this (above)
> eventually.

Did you use dates such as "2017-02-03"?

> > Can you try mozregression without the GUI?
> 
> Sure.  Link for windows D/L?

http://mozilla.github.io/mozregression/install.html#mozregression

It is a bit more involved than the GUI.
(In reply to Marco Castelluccio [:marco] from comment #13)
> (In reply to Codacoder from comment #12)
> > (In reply to Marco Castelluccio [:marco] from comment #11)
> > > (In reply to Codacoder from comment #6)
> > > > I tried to find a regression pointer with mozregression-gui but it crashes
> > > > out with an error asking me to file a bug. Oh the irony ;)
> > > 
> > > Can you paste the error here?
> > 
> > platform: Windows-7-6.1.7601-SP1
> > python: 2.7.11 FROZEN (32bit)
> > mozregui: 0.9.2
> > mozregression: 2.3.2
> > message: IndexError: list index out of range
> > traceback:   File ".\mozregui\bisection.py", line 82, in check_merge
> >   File "..\mozregression\bisector.py", line 241, in handle_merge
> >   File "..\mozregression\build_range.py", line 94, in __getitem__
> > 
> > I tried a few different start/end points, they all do this (above)
> > eventually.
> 
> Did you use dates such as "2017-02-03"?

Yes. See attachment (next post).

> 
> > > Can you try mozregression without the GUI?
> > 
> > Sure.  Link for windows D/L?
> 
> http://mozilla.github.io/mozregression/install.html#mozregression
> 
> It is a bit more involved than the GUI.

I used it ages ago for something else - don't remember where or when and certainly don't have the download available.  If you can do some hand-holding, I'll do the tests.  Your call.
Attached image mozregr.png — 
mozregression-gui bug.
(In reply to Codacoder from comment #14)
> (In reply to Marco Castelluccio [:marco] from comment #13)
> > (In reply to Codacoder from comment #12)
> > > (In reply to Marco Castelluccio [:marco] from comment #11)
> > > > (In reply to Codacoder from comment #6)
> > > > > I tried to find a regression pointer with mozregression-gui but it crashes
> > > > > out with an error asking me to file a bug. Oh the irony ;)
> > > > 
> > > > Can you paste the error here?
> > > 
> > > platform: Windows-7-6.1.7601-SP1
> > > python: 2.7.11 FROZEN (32bit)
> > > mozregui: 0.9.2
> > > mozregression: 2.3.2
> > > message: IndexError: list index out of range
> > > traceback:   File ".\mozregui\bisection.py", line 82, in check_merge
> > >   File "..\mozregression\bisector.py", line 241, in handle_merge
> > >   File "..\mozregression\build_range.py", line 94, in __getitem__
> > > 
> > > I tried a few different start/end points, they all do this (above)
> > > eventually.
> > 
> > Did you use dates such as "2017-02-03"?
> 
> Yes. See attachment (next post).

It looks like an old version of the GUI. Can you try upgrading? (simply remove the old version and download the one)

If the new version doesn't work, we'll try with the CLI version.

BTW, are you sure about the initial range? Are you sure there was no bug on 2017-07-20?
(In reply to Marco Castelluccio [:marco] from comment #16)
> It looks like an old version of the GUI. Can you try upgrading? (simply
> remove the old version and download the one)

Tomorrow.
 
> BTW, are you sure about the initial range? Are you sure there was no bug on
> 2017-07-20?

No, I was winging it just to show you the error in the gui.

But even so, since the bug came, went, then came back again, I'm really not sure what date to use.  If I go by the date this report was created, should I start the day before or the exact same date?  And since I no longer see crash reports (because the bug noted at comment #5 is very likely a different bug that doesn't create reports), should I go by that date? The day before?

Sorry if that's confusing.  Let me just make sure you understand this:  The about:crashes reports up top represent the initial bug.  A few days later, the bug magically disappeared.  A few days later, the bug returned again but with slightly different behavior - the browser hung immediately the drag started and the browser no longer provided about:crashes reports.

So my concern is, if you go by those crash reports, you may in fact be dealing with the wrong data.  

Hope that's clear?

I'll try the new gui tomorrow.
Sorry, I can't find the regression.
Hi Codacoder,
Is this still an issue?

Mozregression tool installation and usage information is attached if the issue is still reproducible.

Thanks for taking time on this.
Flags: needinfo?(codacodercodacoder)
(In reply to Abe - QA (:Abe_LV) from comment #19)
> Created attachment 8900399 [details]
> Mozregression for windows.txt
> 
> Hi Codacoder,
> Is this still an issue?
> 
> Mozregression tool installation and usage information is attached if the
> issue is still reproducible.
> 
> Thanks for taking time on this.

No problem.  Happy to help.

No, hasn't been an issue for over two weeks (few days after my last comment above).  And if it does reappear, I'll now about it right away (I develop in nightly, and tearing off a tab is a part of my current daily workflow).
Flags: needinfo?(codacodercodacoder)
Closing this as incomplete based on comment 20. If anyone can still reproduce it on latest versions, feel free to reopen the bug and provide more information. Thanks
Status: NEW → RESOLVED
Closed: 4 years ago
Resolution: --- → INCOMPLETE
Moving from Core::Untriaged to Core::General https://bugzilla.mozilla.org/show_bug.cgi?id=1407598
Component: Untriaged → General
I filed bug 1419108 for the resurgence which seems to involve accessibility.
Just to say, this bug as originally reported by me in comment 0 and comment 5, has NOT reappeared here.
Ah, I see, it's back in 57.  Tearing off a tab results in a browser unresponsive to mouse and keys.  59 is fine.
You need to log in before you can comment on or make changes to this bug.