Closed Bug 497304 Opened 16 years ago Closed 16 years ago

Crash [@ nsAccessibleWrap::get_accParent(IDispatch**) ]

Categories

(Core :: Disability Access APIs, defect)

1.9.0 Branch
x86
Windows XP
defect
Not set
critical

Tracking

()

VERIFIED FIXED
mozilla1.9.2a1
Tracking Status
blocking1.9.1 --- .3+
status1.9.1 --- .3-fixed

People

(Reporter: MarcoZ, Assigned: MarcoZ)

References

Details

(4 keywords)

Crash Data

Attachments

(1 file)

This is a top crash in the Firefox 3.0.10 release within the last week: http://crash-stats.mozilla.com/report/list?product=Firefox&version=Firefox%3A3.0.10&platform=windows&query_search=signature&query_type=contains&query=Access&date=&range_value=1&range_unit=weeks&do_query=1&signature=nsAccessibleWrap%3A%3Aget_accParent%28IDispatch**%29 Seems to have something to do with tablet PC input, one report mentions returning to the previous page and crashing then. A module named tiptsf.dll is also involved in those of the reports I checked, and according to Microsoft this is a DLL used for the input from the tablet.
Look like we should null check view->GetViewManager()
Attached patch Make sure viewManager is valid (deleted) — Splinter Review
Ginn, is this what you had in mind?
Assignee: nobody → marco.zehe
Status: NEW → ASSIGNED
Attachment #382475 - Flags: review?(ginn.chen)
Attachment #382475 - Flags: review?(ginn.chen) → review?(bolterbugz)
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid Oops, noticed that we're in Windows, not Solaris or Linux. Reassigning review request.
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid r=me with nits: 1. indenting is off 2. I don't know if returning E_FAIL or E_UNEXPECTED is correct. Would be good to understand why/when we don't get a view manager.
Attachment #382475 - Flags: review?(bolterbugz) → review+
(In reply to comment #4) > (From update of attachment 382475 [details] [diff] [review]) > r=me with nits: > 1. indenting is off Thanks! Fixed locally. I switched to NotePad++ and forgot that it should replace tabs with spaces when auto-indenting. > 2. I don't know if returning E_FAIL or E_UNEXPECTED is correct. I tried to be consistent: Further down in the same method there's another instance where we return E_UNEXPECTED. > Would be good to understand why/when we don't get a view manager. Well way down in the call stack, you see that the destructor of viewManager is called, but further up its constructor is never called again. My guess is that tablet pc asks for the accParent at a time where Firefox is taking away one page and replacing it with another. One commenter said he was going back to the previous page when it crashed on him.
Status: ASSIGNED → RESOLVED
Closed: 16 years ago
Resolution: --- → FIXED
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid Requesting approval for 1.9.1.x (but it's not available yet) and 1.9.0.12, was actually found in 1.9.0.10.
Attachment #382475 - Flags: approval1.9.1?
Attachment #382475 - Flags: approval1.9.0.12?
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid I hope you didn't check in with tabs instead of spaces. Please fix your editor settings to prevent that in the future.
Attachment #382475 - Flags: superreview-
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid This isn't looking like a topcrash to us anymore, maybe the software in question got updated? We'll look at this after landing on 1.9.1
Attachment #382475 - Flags: approval1.9.0.12? → approval1.9.0.13?
I corrected the tabs problem locally before checking in, yes! :) We should definitely keep watching out for this and hope those users haven't just given up on Firefox because of this.
Verified fixed in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090618 Minefield/3.6a1pre (.NET CLR 3.5.30729). Patch caused no regressions, and haven't seen any reports of this type creep up on crash-stats in the last week.
Status: RESOLVED → VERIFIED
Attachment #382475 - Flags: superreview-
Attachment #382475 - Flags: approval1.9.0.13?
Attachment #382475 - Flags: approval1.9.0.13+
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid Approved for 1.9.0.13, a=dveditz for release-drivers
Checking in accessible/src/msaa/nsAccessibleWrap.cpp; /cvsroot/mozilla/accessible/src/msaa/nsAccessibleWrap.cpp,v <-- nsAccessibleWrap.cpp new revision: 1.122; previous revision: 1.121 done
Keywords: fixed1.9.0.13
Attachment #382475 - Flags: approval1.9.1.1?
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid We'll get this in Firefox 3.5.3
Attachment #382475 - Flags: approval1.9.1.2? → approval1.9.1.3?
blocking1.9.1: --- → .3+
Comment on attachment 382475 [details] [diff] [review] Make sure viewManager is valid Approved for 1.9.1.3, a=dveditz for release-drivers
Attachment #382475 - Flags: approval1.9.1.3? → approval1.9.1.3+
Severity: major → critical
Target Milestone: --- → mozilla1.9.2a1
Crash Signature: [@ nsAccessibleWrap::get_accParent(IDispatch**) ]
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: