71512 - Assert in GetEditInterface: NS_ENSURE_TRUE(edit,...) failed

Status: VERIFIED WORKSFORME

(Core :: DOM: Events, defect)

Description

[Randell Jesup [:jesup] (needinfo me)]

FreeBSD 4.1 20010308xx

Went to the above URL (from prkjockeys) and clicked on clear.valog (I may have
double-clicked) and got these assertions (see attached GDB backtrace/dumps).

  nsDebug::Assertion (aStr=0x29949145 "NS_ENSURE_TRUE(edit) failed", 
    aExpr=0x29949140 "edit", aFile=0x299467a0 "nsGlobalWindow.cpp", aLine=4403)
    at nsDebug.cpp:189

Comment 1

[Randell Jesup [:jesup] (needinfo me)]

Attachment: GDB backtrace and structure dumps

Comment 2

[clayton]

Please test on mac, windoze and linux.

Comment 3

[Heikki Toivonen (remove -bugzilla when emailing directly)]

I am getting different assertion on yesterday's Linux build. Several of these:

###!!! Break: at file nsXULElement.cpp, line 4284
###!!! ASSERTION: no document: 'mDocument != nsnull', file nsXULElement.cpp,
line 4284

Comment 4

[Heikki Toivonen (remove -bugzilla when emailing directly)]

Tested on Windows with today's build, I get the same assertions I get on Linux.

Reporter, please retest with latest build, and if you will not get the assert
that you reported originally, please mark this worksforme.

Comment 5

[Randell Jesup [:jesup] (needinfo me)]

I get several (6 or 7) mDocument == null assertions (the first is for cmd_undo)
currently.

Perhaps we should transform this into a bug on the mDocument == nsnull
assertions?

If not, go ahead and mark WORKSFORME and file a new bug; otherwise change the
bug title and reassign.

Attaching backtrace for mDocument == null's

Comment 6

[Randell Jesup [:jesup] (needinfo me)]

Attachment: gdb backtrace of mDocument != nsnull asserts

Comment 7

[joki (gone)]

Assertions definitely still there.  Since we're hitting them coming out of 
nsFocusController which (at least according to cvsblame) looks like hyatt's baby 
I'm going to let him look at this.

Comment 8

[David Jensen]

WORKSFORME, Build 2002052306. Tested on Windoze 98, Mandrake 8.2 & OpenBSD 3.1
(earlier build tho)... I know the listed OS is FreeBSD, but then there's
Additional  Comment #2. Unless there's still a problem, why doesn't someone
change the state... unless this bug's very existence has been forgotten - though
it's marked qawanted.

Comment 9

[Derek Schrock]

WORKSFORME build 2002080508 MacOS X 10.1.5

There's the MacOS test... bury the hatchet

Comment 10

[Vladimir Ermakov]

Yes, why dont someone...

Comment 11

[Vladimir Ermakov]

verifying per coments