1309686 - Crash in mozilla::a11y::DocAccessible::UncacheChildrenInSubtree

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[Esther Monchari]

This bug was filed from the Socorro interface and is 
report bp-4dd439a5-1064-416e-9efc-00fee2160928
=============================================================

Crash stated with build id 20160927030200 and is still continuing.

Adding ni on :Alexander and Ehsan since it appears they've worked in that area,  maybe they can help figure out if I placed this in the right component.

Comment 1

[David Bolter [:davidb] (NeedInfo me for attention)]

This is the right component, thanks for the bug report.

Comment 2

[alexander :surkov (:asurkov)]

is regression bug known?

Comment 3

[Marcia Knous [:marcia]]

Adding some keywords to get some traction on finding the regression window.

Comment 4

[BugBot [:suhaib / :marco/ :calixte]]

Crash volume for signature 'mozilla::a11y::DocAccessible::UncacheChildrenInSubtree':
 - nightly (version 52): 29 crashes from 2016-09-19.
 - aurora  (version 51): 80 crashes from 2016-09-19.
 - beta    (version 50): 26 crashes from 2016-09-20.
 - release (version 49): 427 crashes from 2016-09-05.
 - esr     (version 45): 53 crashes from 2016-07-25.

Crash volume on the last weeks (Week N is from 10-17 to 10-23):
            W. N-1  W. N-2  W. N-3  W. N-4
 - nightly       7      13       6       2
 - aurora       28      29      14       2
 - beta          5       7       9       1
 - release     113     134     125      21
 - esr           1       1       0       3

Affected platform: Windows

Crash rank on the last 7 days:
           Browser   Content     Plugin
 - nightly #188      #256
 - aurora  #41
 - beta    #1436
 - release #515
 - esr     #5720

Comment 5

[David Bolter [:davidb] (NeedInfo me for attention)]

Marco have you ever hit this? Would be great to get STR.

Comment 6

[Marco Zehe (:MarcoZ)]

(In reply to David Bolter [:davidb] from comment #5)
> Marco have you ever hit this? Would be great to get STR.
Unfortunately not! I would have either filed it myself if I had, or commented here with a link to my report and a description of the situation this occurred in. :(

Comment 7

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

we are a week away from RC, too late, this is now a wontfix for 50. Also, this is a very low volume crash on Beta50 cycle.

Comment 8

[Gerry Chang [:gchang]]

Hi David,
It seems to me that the crash starts to show from 51.0b6. Can you help find someone to look at this?

Comment 9

[alexander :surkov (:asurkov)]

might have same reasons as bug 1321384 has

I'm curious, if bug 1316789 landing has any correlations with the crash ratio?

Comment 10

[David Bolter [:davidb] (NeedInfo me for attention)]

(In reply to alexander :surkov from comment #9)
> might have same reasons as bug 1321384 has
> 
> I'm curious, if bug 1316789 landing has any correlations with the crash
> ratio?

Yes I think it roughly doubled. Do you have a hypothesis?

Comment 11

[alexander :surkov (:asurkov)]

(In reply to David Bolter [:davidb] from comment #10)
> (In reply to alexander :surkov from comment #9)
> > might have same reasons as bug 1321384 has
> > 
> > I'm curious, if bug 1316789 landing has any correlations with the crash
> > ratio?
> 
> Yes I think it roughly doubled. Do you have a hypothesis?

I would expect it helping rather than aggravating it.

It's interesting the bug was presented in 45.5.1esr only and starting from 50.0.1 (no 46-49). Is there any chance that some changes backported from 50/51 to 45esr increased the crash ratio?

Also 51b/51 crash number increased comparing to 50, which should mean that 51b/51 got some destabializing extras.

Comment 12

[alexander :surkov (:asurkov)]

note, bug 1321384 also starts its story from  50.1.0. It'd be interesting to investigate the list of a11y changes made between 49 and 50 (presumably we've got a crash signature in 50) and between 50 and 51 (crash ratio had increased).

Comment 13

[Justin [:JW_SoftvisionQA]]

In regards to the tracking flags I see that the version is 52 but 51 is shown as affected. Can someone with history of this issue fix that? Thanks

Comment 14

[Marcia Knous [:marcia]]

(In reply to Justin [:JW_SoftvisionQA] from comment #13)
> In regards to the tracking flags I see that the version is 52 but 51 is
> shown as affected. Can someone with history of this issue fix that? Thanks

According to the crash stats signature, http://bit.ly/2jEnv3T, 51 is affected.

Comment 15

[David Bolter [:davidb] (NeedInfo me for attention)]

Still around: https://crash-stats.mozilla.com/report/index/8eeff6d6-cdfa-4b2d-b5e7-9b6542170206

Comment 16

[Marcia Knous [:marcia]]

3232 crashes in the last 7 days on release - this is the #7 overall browser crash.

Comment 17

[alexander :surkov (:asurkov)]

I'm getting certain that it has same roots as bug 1321384 has. Let's keep tracking of these two bugs in a bunch.

Comment 18

[Liz Henry (:lizzard Please n-i to RyanVM, jcristau, or pascal)]

OK, thank.  if that works then we should see this signature drop off in beta 7 later this week.

Comment 19

[Julien Cristau [:jcristau]]

This is still around in b7 and b8, no noticeable drop since previous betas afaict.

Comment 20

[Julien Cristau [:jcristau]]

Not a huge volume and we're getting close to 52 RC → wontfix.

Comment 21

[Liz Henry (:lizzard Please n-i to RyanVM, jcristau, or pascal)]

Wontfix for 53 as well. There are still some crashes on 52 and 53, none so far on 54/55.

Comment 22

[alexander :surkov (:asurkov)]

appears fixed by bug 1363027, landed on moz 54b12 May 25, no new crashes after that, the last one is for 54.0b11 https://crash-stats.mozilla.com/signature/?product=Firefox&signature=mozilla%3A%3Aa11y%3A%3ADocAccessible%3A%3AUncacheChildrenInSubtree&date=%3E%3D2017-05-23T18%3A07%3A00.000Z&date=%3C2017-05-30T18%3A07%3A00.000Z&_columns=date&_columns=product&_columns=version&_columns=build_id&_columns=platform&_columns=reason&_columns=address&_columns=install_time&_sort=-build_id&_sort=-date&page=1