380734 - touching mouse while a page is downloading often crashes SeaMonkey [@ nsContentUtils::ContentIsDescendantOf]

Status: RESOLVED DUPLICATE of bug 380716

(Core :: DOM: UI Events & Focus Handling, defect)

Description

[Kent Paul Dolan]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a5pre) Gecko/20070513 SeaMonkey/1.5a
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a5pre) Gecko/20070513 SeaMonkey/1.5a

Five times in the last 18 hours, at least three of them generating
TalkBack captures, in any way touching the mouse while waiting for
a Google Groups "post news article" operation to complete, crashed
SeaMonkey. This seems to be a more general problem, that pretty
much for any page, moving the mouse, or moving the mouse wheel, or
clicking either mouse button, risks a browser crash if a page is
not yet completely downloaded.

xanthian.


Reproducible: Sometimes

Steps to Reproduce:
1. Use SeaMonkey a lot.
2. Fiddle with the mouse while waiting for a page to finish downloading.
3. Expect a crash "by and by" from this model of behavior.
Actual Results:  
SeaMonkey crashes, sometimes with Talkback, sometimes nerfing Talkback too.


Expected Results:  
Browser is not crashed by stray mouse events while it is busy elsewhere.

This behavior seems independent of how long the browser has been in use,
of what kind of page is downloading, so long as there is a perceptible
download interval, and, unfortunately, is _not_ a "happens every time"
thing, since I rarely let go of the mouse, but only see these crashes
several times a day instead of "with each downloaded page".

I'm on cable broadband, so the speed at which data is arriving may be a
factor.

Question: how does one acquire a "Talkback crash ID"? I don't seem to
receive any such datum with a crash. That term on the bugzilla bug
input form should link to a page explaining how it is to be acquired.

xanthian.

Comment 1

[Mats Palmgren (inactive)]

> how does one acquire a "Talkback crash ID"?

Run talkback.exe
http://kb.mozillazine.org/Quality_Feedback_Agent

Comment 2

[Mats Palmgren (inactive)]

> That term on the bugzilla bug input form should link to a page
> explaining how it is to be acquired.

bug 250604

Comment 3

[Kent Paul Dolan]

(In reply to comment #1)
> > how does one acquire a "Talkback crash ID"?
> 
> Run talkback.exe
> http://kb.mozillazine.org/Quality_Feedback_Agent
> 

Okay, thanks. Here, then, are the four crashes (one more
than I remembered as evoking Talkback out of the five total)
that Talkback caught for this bug:

TB32158568Z, TB32139469M, TB32137438G, TB32133951X

Comment 4

[Kent Paul Dolan]

(In reply to comment #2)
> > That term on the bugzilla bug input form should link to a page
> > explaining how it is to be acquired.

> bug 250604


Thanks, Mats, I added my support to your call for an
easy fix there. [This is a "shoemaker's children"
issue: not having talkback IDs easy for naive reporters
to find and furnish hurts probably the same people who
should be fixing bug 250604 about three years ago.]

xanthian.

Comment 5

[Mats Palmgren (inactive)]

nsContentUtils::ContentIsDescendantOf  [mozilla/content/base/src/nscontentutils.cpp, line 1150]
nsGenericElement::PreHandleEvent  [mozilla/content/base/src/nsgenericelement.cpp, line 1857]
nsEventStateManager::DispatchMouseEvent  [mozilla/content/events/src/nseventstatemanager.cpp, line 2701]
nsEventStateManager::NotifyMouseOver  [mozilla/content/events/src/nseventstatemanager.cpp, line 2822]
nsEventStateManager::GenerateMouseEnterExit  [mozilla/content/events/src/nseventstatemanager.cpp, line 2855]
nsEventStateManager::PreHandleEvent  [mozilla/content/events/src/nseventstatemanager.cpp, line 718]

Comment 6

[Mats Palmgren (inactive)]

Possibly bug 380716?

Comment 7

[Olli Pettay [:smaug][[e-mail]]]

yes

Comment 8

[Kent Paul Dolan]

Thanks for the quick fix, guys, this has been a nasty
one for me, crashing several times a day. Now to wait
_another_ day for the build to propagate.

xanthian.