Closed Bug 1267376 Opened 8 years ago Closed 8 years ago

crash in mozilla::a11y::DocAccessibleParent::RecvShowEvent

Categories

(Core :: Disability Access APIs, defect)

48 Branch
Unspecified
Windows NT
defect
Not set
critical

Tracking

()

RESOLVED FIXED
mozilla49
Tracking Status
firefox47 --- unaffected
firefox48 + fixed
firefox49 + fixed

People

(Reporter: davidb, Assigned: tbsaunde)

Details

(Keywords: crash, regression, topcrash)

Crash Data

Attachments

(1 file)

This bug was filed from the Socorro interface and is 
report bp-37977b44-439d-4c4d-9674-f42f92160425.
=============================================================

More reports here: https://crash-stats.mozilla.com/report/list?product=Firefox&signature=mozilla%3A%3Aa11y%3A%3ADocAccessibleParent%3A%3ARecvShowEvent#tab-reports

Earliest build id seems to be: 20160425030548
(In reply to David Bolter [:davidb] from comment #0)
> Earliest build id seems to be: 20160425030548

Correction: 20160423030220
Suspect caused by recent landings in bug 1262563.
https://crash-stats.mozilla.com/signature/?product=Firefox&release_channel=nightly&platform=Windows&date=%3E%3D2016-04-01&signature=mozilla%3A%3Aa11y%3A%3ADocAccessibleParent%3A%3ARecvShowEvent#aggregations
is a query for all crashes with this signature in nightly channel since April 1.

Agrees with the comment 1 correction to 20160423030220 which suggests a regression range (computed from http://dbaron.org/mozilla/crashes-by-build ) as https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=0891f0fa044cba28024849803e170ed7700e01e0&tochange=37f04460ddb76d6ef4e7c32a8a6b2fbc44cb8776 , which agrees with comment 2's suspicion.


Crashes in mozilla::a11y::AccessibleWrap::GetChildIDFor :
https://crash-stats.mozilla.com/signature/?product=Firefox&release_channel=nightly&platform=Windows&date=%3E%3D2016-04-01&signature=mozilla%3A%3Aa11y%3A%3AAccessibleWrap%3A%3AGetChildIDFor#aggregations
seem to cover roughly the same set of builds (though a little less frequent).  Does it seem like this bug can cover them as well, or should a separate bug be filed?
Keywords: topcrash
> Crashes in mozilla::a11y::AccessibleWrap::GetChildIDFor :
> https://crash-stats.mozilla.com/signature/
> ?product=Firefox&release_channel=nightly&platform=Windows&date=%3E%3D2016-04-
> 01&signature=mozilla%3A%3Aa11y%3A%3AAccessibleWrap%3A%3AGetChildIDFor#aggrega
> tions
> seem to cover roughly the same set of builds (though a little less
> frequent).  Does it seem like this bug can cover them as well, or should a
> separate bug be filed?

I think the theory is the object passed to FireWinEvent() is no good, which would explain that crash too, so probably no point in another bug atm.
Attachment #8745570 - Flags: review?(dbolter) → review+
[Tracking Requested - why for this release]: Crash bug caused by bug 1262563, which landed on 48. Need to backport fix to Aurora. Requesting tracking to make sure this doesn't get overlooked.
Adding [@ mozilla::a11y::AccessibleWrap::GetChildIDFor ] signature for now so it gets picked up in crash-stats per Comment 5.
Crash Signature: [@ mozilla::a11y::DocAccessibleParent::RecvShowEvent] → [@ mozilla::a11y::DocAccessibleParent::RecvShowEvent] [@ mozilla::a11y::AccessibleWrap::GetChildIDFor]
Assignee: nobody → tbsaunde+mozbugs
needinfo? to remember to make aurora approval requests
Flags: needinfo?(tbsaunde+mozbugs)
Tracking for 48/49, recent regression
https://hg.mozilla.org/mozilla-central/rev/7f233aaf281d
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla49
I was (trying to) add a comment in a Google Doc, and Firefox nightly completely crashed on me. My nightly is current (May 2)
BTW, I run Windows 10 on a Lenovo X1 Carbon Touch. 

This GDoc is unusable (as they all now, for the most part, if I try making changes) and as a result, I have to do _all_ my GDocs work in Chrome :/
Karen, are you sure this is this bug that you're crashing in? Because this was fixed for the 04-29 or 04-30 Nightly build. And it should affect e10s only. Do you have the link to a report?
Flags: needinfo?(krudnitski)
This is the crash ID (https://crash-stats.mozilla.com/report/index/86967a18-22e1-402e-84f3-efe842160503) , and it automatically linked it to this bug report, so I just assumed it was smarter than me!
Flags: needinfo?(krudnitski)
(In reply to Karen Rudnitski [:kar] from comment #15)
> This is the crash ID
> (https://crash-stats.mozilla.com/report/index/86967a18-22e1-402e-84f3-
> efe842160503) , and it automatically linked it to this bug report, so I just
> assumed it was smarter than me!

Oh good that crash was on build ID 20160428030218 -- if you can please apply the latest nightly update and report back it would be great to confirm that we fixed this for you as well.
For the record, I checked my Firefox was updated at the time I submitted this report. But since you mentioned the build ID, I went back and have now, indeed, found a new version.

It's installed, and I've left a few comments on that doc and so far, so good. I'll open a few more docs over the course of the day and see what happens.
Comment on attachment 8745570 [details] [diff] [review]
bail out of RecvShowEvent() early if we failed to add children

Approval Request Comment
[Feature/regressing bug #]: 1262563
[User impact if declined]: crashes
[Describe test coverage new/current, TreeHerder]:  its been in m-c for a couple days.
[Risks and why]:  low, it just stops running code added in the bug that caused this in some cases.
[String/UUID change made/needed]: none
Flags: needinfo?(tbsaunde+mozbugs)
Attachment #8745570 - Flags: approval-mozilla-aurora?
Comment on attachment 8745570 [details] [diff] [review]
bail out of RecvShowEvent() early if we failed to add children

Crash fix, recent regression from 48, tested in m-c. Please uplift to aurora.
Attachment #8745570 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
Version: unspecified → 48 Branch
You need to log in before you can comment on or make changes to this bug.