931794 - startup crash in mozilla::dom::WindowBinding::get_content with Twitter Disconnect, Facebook Disconnect, Google Disconnect, Chatzilla

Status: VERIFIED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Daniel Holbert [:dholbert]]

This bug was filed from the Socorro interface and is 
report bp-4f9b57fa-e48a-4ce4-9f6d-068d82131028.
=============================================================

STR:
 1. Start latest nightly, with fresh profile to be on the safe side
 2. Install Twitter Disconnect from here:
   https://addons.mozilla.org/en-US/firefox/addon/twdc/
 3. Restart Firefox to complete installation.

ACTUAL RESULTS: Startup crash (unable to start firefox, except in safe mode).

Sample crash reports:
 bp-4f9b57fa-e48a-4ce4-9f6d-068d82131028
 bp-f55a1398-1f02-4bc4-b3d1-b1b512131028
 bp-a6de9a2f-7eac-464d-a4d9-45c982131028
 bp-e89a57cd-d7d4-4c2d-ac85-46a772131028

This is with today's nightly [27.0a1 (2013-10-28)] and Twitter Disconnect 2.1.2

Comment 1

[Daniel Holbert [:dholbert]]

I also hit this with Facebook Disconnect:
 https://addons.mozilla.org/en-us/firefox/addon/fbdc/
 bp-97146e4f-a04d-45e2-9b8d-20cd02131028
and Google Disconnect:
 https://addons.mozilla.org/En-us/firefox/addon/gdc/
 bp-d78fdbc7-3a5b-497c-b52c-d8de12131028

(In each case, the add-on in question is the only one enabled)

Comment 2

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

I'm going to guess that this is a DOM bindings issue.

Comment 3

[Daniel Holbert [:dholbert]]

This is new in latest nightly, based on my testing.

Regression range, given that fact:
  http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=a80dce1126db&tochange=4646259ab62d

Maybe a regression from bug 918345?

Comment 4

[Olli Pettay [:smaug][[e-mail]]]

Very likely.

Comment 5

[Daniel Holbert [:dholbert]]

Yup, I confirmed that http://hg.mozilla.org/integration/mozilla-inbound/rev/508288a2b62c (for bug 918345) is the first affected cset.

Comment 6

[Boris Zbarsky [:bzbarsky]]

These are all null-derefs in js::GetObjectCompartment.

I wonder whether GetContent() is returning null (and failing some asserts on its toObject() calls in the process; presumably this is the case when treeOwner->GetContentWindow() returns null?  We should probably mark this nullable in the idl and change that last return to "return val.toObjectOrNull()".

Comment 7

[Daniel Holbert [:dholbert]]

Setting "Tracking" flag, to get this on rel drivers' radar. We absolutely should not ship this startup crash to Aurora users during the train-switch that's happening in the next ~24 hours.

Comment 8

[bhavana bajaj [:bajaj]]

Tracking as we do not want to enable aurora updates without this bug being resolved.

Comment 9

[Peter Van der Beken [:peterv]]

Attachment: v1

Comment 10

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 823523 [details] [diff] [review]
v1

r=me

Comment 11

[B.J. Herbison]

Chatzilla is also causes the issue (new profile, only installed Chatzilla).
bp-dcd4b72a-b402-42cb-8173-ce3242131029
bp-46bff72e-daea-4719-98ba-515052131029

Comment 12

[Peter Van der Beken [:peterv]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/a4455e4749a0

Comment 13

[Peter Van der Beken [:peterv]]

Comment on attachment 823523 [details] [diff] [review]
v1

[Approval Request Comment]
Bug caused by (feature/regressing bug #): bug 918345
User impact if declined: crashes
Testing completed (on m-c, etc.): landed on mozilla-inboud
Risk to taking this patch (and alternatives if risky): low risk, just deals with a null pointer
String or IDL/UUID changes made by this patch: none

Comment 14

[Peter Van der Beken [:peterv]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/f31ef270c830

Comment 15

[u279076]

Flagging this as a topcrash since it's the #1 on Nightly by far. I'll monitor crashstats over the next couple days to ensure this is fixed. When did this land on mozilla-central?

Comment 16

[Ryan VanderMeulen [:RyanVM]]

It hasn't been merged yet - it's held up by the current trunk closure. I set the status flag for firefox28 mainly so it was clear that it didn't need to land on any other branches at this point. Sorry for any confusion that caused.

Comment 17

[Robert Kaiser]

Yes, having 28 set as fixed when it isn't is indeed confusing. 26 and 27 are marked correctly anyhow, why not leave 28 as "affected" until this lands?

Comment 18

[Ryan VanderMeulen [:RyanVM]]

Because I find that often people forget to set it and mcMerge doesn't do it automatically unless it's tracking+.

Comment 19

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/a4455e4749a0

Comment 20

[Daniel Holbert [:dholbert]]

I'm hitting something like this bug in recent nightlies on Windows, too (after this patch landed) - see bug 934203.

Comment 21

[Alice0775 White]

There are 2 regressions
#1 Crash [@ mozilla::dom::WindowBinding::get_content ]
   http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=a80dce1126db&tochange=4646259ab62d
   Troggered by Bug 918345
#2 Crash [@ JS_WrapValue(JSContext*, JS::MutableHandle<JS::Value>) ]
   http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=f0d363d72753&tochange=abe6790a5dd8
   Triggered by bug 931794

Regression window(m-c Nightly)
Good:
http://hg.mozilla.org/mozilla-central/rev/a80dce1126db
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20131027030205

#1 Bad [@ mozilla::dom::WindowBinding::get_content ]:
bp-40b8c650-6ac1-4c7d-b949-62cc12131103
http://hg.mozilla.org/mozilla-central/rev/4646259ab62d
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20131028030205
bp-663345c9-1ec0-41f1-b208-47de32131103
http://hg.mozilla.org/mozilla-central/rev/518f5bff0ae4
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:28.0) Gecko/20100101 Firefox/28.0 ID:20131029030201
bp-24cd6872-8622-4d2f-b7af-66b872131103
http://hg.mozilla.org/mozilla-central/rev/829d7bef8b0a
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:28.0) Gecko/20100101 Firefox/28.0 ID:20131030030201
bp-8b4cd97e-4803-4d2e-bf64-5dcf72131103
http://hg.mozilla.org/mozilla-central/rev/f0d363d72753
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:28.0) Gecko/20100101 Firefox/28.0 ID:20131031030203

#2 Bad [@ JS_WrapValue(JSContext*, JS::MutableHandle<JS::Value>) ]:
bp-a0b80ddd-4423-43f9-b113-99ed32131103
http://hg.mozilla.org/mozilla-central/rev/abe6790a5dd8
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:28.0) Gecko/20100101 Firefox/28.0 ID:20131101030205
bp-afa204a4-00be-467e-a998-7d6ff2131103
http://hg.mozilla.org/mozilla-central/rev/2f7593f81351
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:28.0) Gecko/20100101 Firefox/28.0 ID:20131102030205
bp-ad99948d-f32f-4a6d-b6fc-d0d0b2131103
http://hg.mozilla.org/mozilla-central/rev/47c8e9b16918
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:28.0) Gecko/20100101 Firefox/28.0 ID:20131103030204

Comment 23

[Ada [:adalucinet]]

Mozilla/5.0 (X11; Linux i686; rv:27.0) Gecko/20100101 Firefox/27.0
Mozilla/5.0 (X11; Linux x86_64; rv:27.0) Gecko/20100101 Firefox/27.0

Reproduced this issue with Nightly from 2013-10-28 and on Aurora from 2013-11-05 I've encountered the startup crash from bug 934203.
Verified as fixed on latest Aurora (Build ID: 20131106004000): no crash when restarting Firefox to complete the installation for Twitter Disconnect 2.1.2, Facebook Disconnect  2.1.3 and Google Disconnect  2.4.2.

Comment 24

[Tracy Walker [:tracy]]

Per crash stats, there are no reports of this crash with recent Nightly builds.