1335522 - Tab crashes on dropbox.com on Mac OSX 10.9 and 10.10

Status: RESOLVED DUPLICATE of bug 1331683

(Core :: Graphics: Text, defect, P3)

Description

[yusuhail]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:54.0) Gecko/20100101 Firefox/54.0
Build ID: 20170131030205

Steps to reproduce:

I started Firefox Nightly 54.0a1 (2017-01-31) (64-bit) in Safe Mode. I went to www.dropbox.com


Actual results:

In a couple of seconds, the tab crashes and this is what I see: 
Gah. Your tab just crashed.
We can help you!

Choose Restore This Tab to reload page content.



Expected results:

The website should have worked.

Comment 1

[Tooru Fujisawa [:arai]]

Can you provide crash report ID for the crash?
  https://support.mozilla.org/en-US/kb/mozillacrashreporter

Comment 2

[yusuhail]

Attachment: Screen Shot 2017-01-31 at 5.43.48 PM.png

screenshot of the tab

Comment 3

[yusuhail]

So this is just the tab crashing rather than Firefox crashing. I know how to get the Firefox crash report, but how do I get the tab crashing report. about:crashes doesn't show any new crashes and the crash dialog box doesn't come up. Firefox remains functional, its only the specific tab that displays the message. I've attached the screenshot. Please let me know how to get the crash report for the tab, or if there is any telemetry info I can provide.

Comment 4

[Loic]

I can't reproduce the crash with the latest Nightly on Win 7.

Is it reproducible at 100% on your machine?

Comment 5

[yusuhail]

Yes, it is happening every time I try it. I updated to today's Nightly and the tab crashed again. I'd like to help ... is there data from the about:telemetry page that I can attach to help diagnose this? As I mentioned, I started Nightly in Safe Mode.

Comment 6

[Ciprian Muresan [:cmuresan], Formerly of Ecosystem QA - please needinfo @mcoman for things regarding the team]

[Tracking Requested - why for this release]:

User Agent  Mozilla/5.0 (Macintosh; Intel Mac OS X 10.10; rv:54.0) Gecko/20100101 Firefox/54.0
Build ID  20170202030211

I was able to reproduce the issue on the latest Nightly (54.0a1), but not on the latest Firefox release (51.0.1).
The issue is reproducible on Mac OS 10.10 and 10.9, but not on 10.11 and 10.12

Last good revision: 44ab9ab869d54feae16cd084cafe36b5595833c4
First bad revision: 273ce98bb9e831b70d9ce2b34590859ea2ce80b8
Pushlog:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=44ab9ab869d54feae16cd084cafe36b5595833c4&tochange=273ce98bb9e831b70d9ce2b34590859ea2ce80b8

Looks like the following bug has the changes which introduced the regression:
https://bugzilla.mozilla.org/show_bug.cgi?id=1321031

Jonathan, can you please take a look at this?

Had to turn off e10s to get a crash to be reported:
10.9  - bp-71bebc44-fdac-4971-8c3a-583c02170203
10.10 - bp-c9f0928b-18b7-484e-8337-9246d2170203

Comment 7

[Loic]

[Tracking Requested - why for this release]:

Comment 8

[Jonathan Kew [:jfkthame]]

This looks like it could be a dupe of bug 1331683. Ciprian, could you please test the try build from bug 1331683 comment 10, and confirm whether you still see the crashes? (And if so, are they the _same_ crashes or different?) Thanks!

Comment 9

[Ciprian Muresan [:cmuresan], Formerly of Ecosystem QA - please needinfo @mcoman for things regarding the team]

I tested the try build on Mac 10.10 with both scenarios and they crash with the same signature I got in comment 6 [@ libsystem_kernel.dylib@0x16286 ] 

It's just a bit different than the one from bug 1331683 comment 0 [@ libsystem_kernel.dylib@0x15866 ].

Comment 10

[Marcia Knous [:marcia]]

Tracking 53/54+ for this regression.

Comment 11

[yusuhail]

The crash signature for e10s window in the latest build seems to have changed. Here is the latest [crash report from today's build](https://crash-stats.mozilla.com/report/index/ed12326e-8a1b-45ce-9ae1-5e0372170207).

Comment 12

[Andrew Overholt [:overholt]]

Jonathan, needinfoing you just in case you didn't see comment 9.

Comment 13

[Jonathan Kew [:jfkthame]]

Yeah, I did see it, thanks. I've finally gotten around to setting up a dev environment on an old (and painfully slow!) machine running 10.10, so I should be able to reproduce and debug this properly now; hope to have a new patch shortly.

Comment 14

[Jonathan Kew [:jfkthame]]

See bug 1331683 comment 14 for a new try build that should fix this. Testing appreciated (once the build is ready) - thanks!

Comment 15

[Ciprian Muresan [:cmuresan], Formerly of Ecosystem QA - please needinfo @mcoman for things regarding the team]

Tested the new try build on Mac 10.10 and 10.9 with both scenarios and the crash is no longer reproducible.

I think we can close this one as a duplicate of bug 1331683. 
I don't see why we should track the same issue in two bugs. Jonathan, what do you think?

Comment 16

[Jonathan Kew [:jfkthame]]

Agreed; I was pretty sure this would turn out to be a duplicate, but I wanted to confirm that the try build fixed the issue for multiple reporters before actually closing it.

Comment 17

[Liz Henry (:lizzard)]

Fixed in 53/54 in the duplicate bug.