310646 - Crash [@ InternalColor2Index() ] when scrolling in recipient list

Status: RESOLVED WORKSFORME

(Thunderbird :: Message Compose Window, defect)

Description

[Dave Miller [:justdave]]

Steps to reproduce:

1) Open an email with a large CC list.
2) Click Reply All
3) Position the mouse over the recipient list
4) use the scroll wheel to scroll down
5) scroll up
6) scroll down again
7) poof.

Using the scroll bar does not seem to affect it, it only happens when you use
the scroll wheel.

Relevant talkback IDs:

2005092906: TB9906699Z TB9906647H
2005093005: TB9926822Z TB9927354H

Comment 1

[Dave Miller [:justdave]]

Attachment: CrashReporter log

CrashReporter crash log, since it's sometimes more useful than the talkback
stack traces.

Comment 2

[Mark Mentovai]

Couldn't reproduce on 10.4.2, which Dave is also running.  My QuickDraw library
is newer than his.  The QD library was updated in Security Update 2005-8.

Comment 3

[Mark Mentovai]

Dave also managed to crash this at TB9930320Y.  I am now able to reproduce. 
I'll investigate.

Comment 4

[Mark Mentovai]

Attachment: Some stacks

Comment 5

[Mark Mentovai]

This was broken at least as far back as 091507.  There are a variety of
different stack traces that this will crash with.  Something is being destroyed.
 Looks like the frames that are scrolled out of view.

The build IDs aren't strongly related to the crashes, they're only the builds
that produced certain crashes.

100105, 092209, and 091507 crash in DispatchEventToHandlers .
nsMacWindow::ScrollEventHandler . SetGWorld

092805 is an interesting crash in DispatchEventToHandlers .
nsMacWindow::ScrollEventHandler .. nsListBoxBodyFrame::InternalPositionChanged .
nsListBoxBodyFrame::DestroyRows .. nsFrameList::DestroyFrames .
nsContainerFrame::Destroy . nsFrame::Destroy . nsView::~nsView .
nsView::DropMouseGrabbing

092607 and 091905 are also interesting, similar to the above crash but after the
nsView destructor, nsBaseWidget::Release . nsMacWindow::~nsMacWindow . (system
window destruction functions)

Comment 6

[Adam Guthrie]

*** Bug 317873 has been marked as a duplicate of this bug. ***

Comment 7

[Wayne Mery (:wsmwk)]

Dave, do you still see this?
bug 331937, which doesn't seem related, is only bug I find with SetPortRGBForeColor in the stack (SetPortRGBForeColor is second in your stack)

Comment 8

[Dave Miller [:justdave]]

nope.  And I remember seeing it listed as getting fixed in the Rumbling Edge report at some point, too, so was this supposed to be a dupe?

Comment 9

[Wayne Mery (:wsmwk)]

dunno about the dupe.
=> WFM - someone can look for the dupe if they wish