684599 - crash nsDeviceMotion::DeviceMotionChanged

Status: VERIFIED FIXED

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

Description

[Dirkjan Ochtman (:djc)]

This bug was filed from the Socorro interface and is 
report bp-e3071077-b38d-4f2a-a7d4-a9dbe2110904 .
============================================================= 

I've had this crash 4 times over two days; this is the first time I've had it. It seems to be related to scrolling events. I'm running OS X Lion 10.7.1 and latest Aurora.

https://crash-stats.mozilla.com/report/index/bp-6062067e-13e5-41d0-b33d-8f29c2110903
https://crash-stats.mozilla.com/report/index/bp-1a4c29fb-dad6-48e3-8872-a956b2110904
http://crash-stats.mozilla.com/report/index/bp-a9bee1df-4d98-4315-8b27-4b5242110903

(The associated bug states that everything after that fix should be considered separate, so here I am.)

Comment 2

[Dirkjan Ochtman (:djc)]

kbrosnan, did you have a particular reason to ignore bug 673472 comment 15?

https://bugzilla.mozilla.org/show_bug.cgi?id=673472#c15

Comment 3

[Josh Matthews [:jdm]]

The buildid says it's from a Sept 2nd build, so this looks like a crash that needs addressing. Dirkjan, any further details about when this reproduces?

Comment 4

[Dirkjan Ochtman (:djc)]

I had a scenario to reproduce after each crash:

1. Restore session (with 4 app tabs and 4 normal tabs)
2. Open new tab
3. In new tab, google for simplejson
4. Click link to http://undefined.org/python/
5. Try to hit the github link under simplejson
6. Crash before getting to click the link

However, at some point, instead of restoring the session, I started clean, opened up all my (app + normal) tabs again, and now it seems to work without crashing.

My previous instance of this had a similar way of reoccurring after the crash, when I tried to do the exact same thing. Ah, see the comment I left on the 606* crash, that also had something to do with github (though in this case I didn't even get to github before crashing). So I guess the first time I had it was not yesterday but this morning; should be a recent regression.

Comment 5

[Josh Matthews [:jdm]]

Thanks for the hint. I was able to reproduce it by opening up two tabs to the github 404 page, refreshing them both, and closing them.

Comment 6

[Josh Matthews [:jdm]]

This is a regression from bug 678818.

Comment 7

[Josh Matthews [:jdm]]

Attachment: Avoid null dereference when chacking if a window's in the background.

This code just keeps on giving, doesn't it?

Comment 8

[Josh Matthews [:jdm]]

http://hg.mozilla.org/integration/mozilla-inbound/rev/81ebc148aad3

Comment 9

[Dirkjan Ochtman (:djc)]

Should there be some flag set to make sure this goes into Aurora, or does that only happen after landing onto mozilla-central?

Comment 10

[Doug Turner (:dougt)]

smaug, how can the outer window ever be null here?  I thought inners always have outers?

jdm - yes, giving and giving and giving.

Comment 11

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

I hit this today, and interesting it didn't trigger breakpad (I got the MS dialog and dropped into the debugger).

Comment 12

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

http://hg.mozilla.org/mozilla-central/rev/81ebc148aad3

Comment 14

[Robert Kaiser]

Flagging for 8 as it appears in crash stats there. I know approval has been requested already, I just don't want to lose track of it. :)

Comment 15

[Josh Matthews [:jdm]]

Can someone push this to aurora and beta for me, please?

Comment 16

[Mark Finkle (:mfinkle) (use needinfo?)]

(In reply to Josh Matthews [:jdm] from comment #15)
> Can someone push this to aurora and beta for me, please?

Pushing

Comment 17

[juan becerra [:juanb]]

We'll need to verify this on aurora and beta when it lands. Use the information in comment #4 and #5 to verify.

Comment 18

[LegNeato]

This looks to have landed everywhere, marking as fixed everywhere.

Comment 19

[Vlad [QA]]

Setting resolution to Verified Fixed on:
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:7.0) Gecko/20100101 Firefox/7.0
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:8.0a2) Gecko/20110921 Firefox/8.0a2
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:9.0a1) Gecko/20110922 Firefox/9.0a1

I've verified this following the steps from comment4 and I forced quit firefox from the Activity Monitor to get the Retore Tab. On all three channels I got no crash.