967087 - crash in js::jit::IonContext::~IonContext()

Status: RESOLVED WORKSFORME

(Core :: JavaScript Engine, defect)

Description

[Richard Newman [:rnewman]]

This bug was filed from the Socorro interface and is 
report bp-b64c9103-dc01-480f-8375-0a9122140203.
=============================================================

Another pile of reproducible crashes in the JS engine on Wired.com. 

On phone, so can't easily grab a reference to the last one.

Comment 1

[Richard Newman [:rnewman]]

Bug 950388 is the other crasher bug on Wired.com.

Comment 2

[Brad Lassey [:blassey] (use needinfo?)]

Naveed, got someone to throw at this?

Comment 3

[Jan de Mooij [:jandem]]

This is ARM only. Marty/Nicolas can one of you reproduce this?

Comment 4

[Richard Newman [:rnewman]]

This is tracking 29. Who's available to look into this?

Comment 5

[Jan de Mooij [:jandem]]

(In reply to Richard Newman [:rnewman] from comment #4)
> This is tracking 29. Who's available to look into this?

Is this still a problem for you? I don't see it on crash-stats for 29?

Comment 6

[Richard Newman [:rnewman]]

Looks like it's not a topcrash, and I can't repro with 31 on Wired's current site, but here's a 29 report from two days ago:

https://crash-stats.mozilla.com/report/index/cbed1a11-0652-4664-867d-617ce2140326

Comment 7

[Jan de Mooij [:jandem]]

(In reply to Richard Newman [:rnewman] from comment #6)
> Looks like it's not a topcrash, and I can't repro with 31 on Wired's current
> site, but here's a 29 report from two days ago:

OK, assigning to Marty then.

Comment 8

[Brad Lassey [:blassey] (use needinfo?)]

Naveed, coming up on the last week of beta for 29

Comment 9

[Marty Rosenberg [:mjrosenb]]

Can someone grab a url out of the crash reporter? I tried for a while on my htc one, and my galaxy tab, with no crashes.

Comment 10

[Brad Lassey [:blassey] (use needinfo?)]

Kevin, can you get the URLs from the crash reports?

Comment 11

[Robert Kaiser]

(In reply to Brad Lassey [:blassey] (use needinfo?) from comment #10)
> Kevin, can you get the URLs from the crash reports?

We do that via setting the needURLs keyword.

Comment 12

[Kevin Brosnan [Ex-Mozilla]]

Crash URL was wyciwyg://23/http://www.wired.com/autopia/2014/02/mulholland-the-snake/ this crash signature is quite uncommon 1 crash in the last 7 days.

Comment 13

[Brad Lassey [:blassey] (use needinfo?)]

(In reply to Robert Kaiser (:[e-mail]) from comment #11)
> (In reply to Brad Lassey [:blassey] (use needinfo?) from comment #10)
> > Kevin, can you get the URLs from the crash reports?
> 
> We do that via setting the needURLs keyword.

thanks for the knowledge.

Marty, can you reproduce now?

Comment 14

[Marty Rosenberg [:mjrosenb]]

I'll look into this this weekend.

Comment 15

[Naveed Ihsanullah [:naveed]]

followed up with Marty. This is at the top of his queue for next week.

Comment 16

[Marty Rosenberg [:mjrosenb]]

(In reply to Kevin Brosnan [:kbrosnan] from comment #12)
> Crash URL was
> wyciwyg://23/http://www.wired.com/autopia/2014/02/mulholland-the-snake/ this
> crash signature is quite uncommon 1 crash in the last 7 days.

I tried to reproduce this with every version of fennec I could find, and no luck.  Is there anything else that has crashed more than once in a week?

Comment 17

[Kevin Brosnan [Ex-Mozilla]]

Crash rate is so low that this is being removed from tracking.

Comment 18

[Wayne Mery (:wsmwk)]

I propose to close this...

No builds newer than 20141218164527 have sig containing js::jit::IonContext::~IonContext 
bp-85315802-244b-4639-a3b9-a8ae92150510	2015-05-10 17:54:55 	
@0x0 | NP_Shutdown | NP_Shutdown | arena_dalloc | NP_Shutdown | js::jit::IonContext::~IonContext() 
FennecAndroid 	34.0.1 	20141218164527 

any objection to closing bug 950388, which also has no crashes?