78102 - Attempt to decrement focus controller's suppression when no suppression active!

Status: RESOLVED INACTIVE

(Core :: DOM: Core & HTML, defect, P5)

Description

[timeless]

###!!! ASSERTION: Attempt to decrement focus controller's suppression when no 
suppression active!: 'PR_FALSE', file 
/tmp/mozilla/dom/src/base/nsFocusController.cpp, line 411
###!!! Break: at file /tmp/mozilla/dom/src/base/nsFocusController.cpp, line 411

SunOS 5.7 Generic_106541-15 sun4u sparc SUNW,Ultra-2
mozilla cvs build for xlib from this morning.

This happens a few seconds after i type a url into the address line and press 
enter.
{steps to reproduce: build xlib mozilla, run mozilla, enter url into location 
bar}
Opening file cookperm.txt failed
Opening file cookperm.txt failed
Disabling View Source StyleSheet
Creating a new nsIToolkit!
Enabling Quirk StyleSheet
/
8x{
###!!! ASSERTION: nsImageXlib::DrawToImage not yet implemented: 'Not Reached', 
file /tmp/mozilla/gfx
/src/xlib/nsImageXlib.cpp, line 1093
###!!! Break: at file /tmp/mozilla/gfx/src/xlib/nsImageXlib.cpp, line 1093
}
<your assertion here>

This does not happen if i use file>open location.

Comment 1

[Dan Rosen]

This is hyatt's realm -- hyatt, I think somebody's not playing nice with you.
Either that or you just shouldn't be asserting here.

Comment 2

[Mike Kaply [:mkaply]]

We get this same assertion repeatedly on OS/2

Comment 3

[timeless]

cc urlbar people, although since both instances have a urlbar widget it doesn't 
make sense. cc a window person, although it'd be more reasonable to cc if the 
assert happened in the other case.

Comment 4

[Prashant Desale]

Updating QA contact to Shivakiran Tummala.

Comment 5

[Mike Kaply [:mkaply]]

This one happens CONSTANTLY on OS/2. It makes debug builds unusable.

Do we have any idea what this is?

Adding saari to the list since he knows focus issues.

Comment 6

[timeless]

come to think of it, this happens constantly on my 092w32branch build. the bad 
thing is if i'm typing i can accidentally trigger abort...

someone claimed this might be the reason for most of the focus issues (windows 
popping to front) according to them the assert appears when windows would 
normally pop to front. of course w32debug instead of them popping to front, i 
get an assert for them that pops to front -- fun.

I suppose we should change this to all/all ...

Comment 7

[Mike Kaply [:mkaply]]

I can get this reasonably consistently on Windows.

Type in www.yahoo.com in the URL bar, then immediately switch to another window.

Assert.

We REALLY need to fix this or get rid of this assert.

Comment 8

[David Hyatt]

I think this has been fixed, but over to the focus team for verification.

Comment 9

[Roland Mainz]

Can anyone please point me to the "fix" if there is one ?
Xlib toolkit still suffers from this problem (output from a 2001-09-19-08-trunk
Xlib toolkit build):
-- snip --
Disabling Quirk StyleSheet Enabling Quirk StyleSheet
###!!! ASSERTION: Attempt to decrement focus controller's suppression when no
suppression active!
: 'PR_FALSE', file
../../../../../../src/2001-09-19-08-trunk/mozilla/dom/src/base/nsFocusController.cpp,
line 416
###!!! Break: at file
../../../../../../src/2001-09-19-08-trunk/mozilla/dom/src/base/nsFocusController.cpp,
line 416
Document http://www.mozilla.org/ loaded successfully
-- snip --

Comment 10

[Mike Kaply [:mkaply]]

I have more info on this at least in the OS/2 case.

My test is simply starting Mozilla and going to Yahoo.

What I noticed in this case is that OS/2 gets three focus messages that Windows 
doesn't for Activation Suppression. An increment and two decrements.

The Activation Suppression increment is coming from:
nsWebShellWindow::HandleEvent for an NS_GOTFOCUS message

Both Activation Suppression decrements are coming from:
nsEventStateManager::PreHandleEvent for an NS_ACTIVATE message.
The while loop at:

http://lxr.mozilla.org/seamonkey/source/content/events/src/nsEventStateManager.c
pp#630

Is being executed twice causing both decrements.

So there are two issues.

1. Why are we getting Activation Suppression that windows isn't?

2. Why are we getting two of the same activation?

I'm copying hyatt because according to the DEBUG_ stuff, he did a lot of this 
stuff.

Comment 11

[Judson Valeski]

this might be gone w/ the landing of 122462, can someone who can repro this try
again w/ a trunk build?

Comment 12

[Mike Kaply [:mkaply]]

Still happens on OS/2 - those changes are definitely there.

Comment 13

[Roland Mainz]

Still happens with Xlib toolkit, too... ;-(

Comment 14

[Mike Kaply [:mkaply]]

I just rewrote the OS/2 focus code so it follows EXACTLY the same messaging as 
Windows and it fixed this assert as well as many other focus issues.

My recommendation is that any platform experiencing this problem use the windows 
build to do a log of a very specific focus case (mine was just bringing up the 
browse and closing it)

Once I got the OS/2 browser matching those messages exactly, every other 
testcase I tried worked.

Comment 15

[Bob Clary [:bc] (inactive)]

Test case from DOM HTML 1 test suite

Does not ASSERT
http://dom-ts.bclary.com/build/ecmascript/level1/html/files/applet.html

Does ASSERT http://dom-ts.bclary.com/build/ecmascript/level1/html/files/applet.xhtml

Does Assert http://dom-ts.bclary.com/build/ecmascript/level1/html/files/applet.xml

###!!! ASSERTION: Attempt to decrement focus controller's suppression when no
suppression active!
: 'PR_FALSE', file
c:/work/mozilla_source/trunk/mozilla/dom/src/base/nsFocusController.cpp, line 457

stack

NTDLL! 77f75a58()
nsDebugImpl::Assertion(nsDebugImpl * const 0x002e6d80, const char * 0x016cbc10,
const char * 0x016cbc00, const char * 0x016cbbb8, int 457) line 272
nsDebug::Assertion(const char * 0x016cbc10, const char * 0x016cbc00, const char
* 0x016cbbb8, int 457) line 109
nsFocusController::SetSuppressFocus(nsFocusController * const 0x03924758, int 0,
const char * 0x015fdc64) line 457 + 35 bytes
PresShell::UnsuppressAndInvalidate() line 4830
PresShell::ProcessReflowCommands(int 1) line 6421
ReflowEvent::HandleEvent() line 6178
HandlePLEvent(ReflowEvent * 0x03a231a0) line 6192
PL_HandleEvent(PLEvent * 0x03a231a0) line 673 + 10 bytes
PL_ProcessPendingEvents(PLEventQueue * 0x009bd7e8) line 608 + 9 bytes
_md_EventReceiverProc(HWND__ * 0x00040110, unsigned int 49376, unsigned int 0,
long 10213352) line 1414 + 9 bytes
USER32! 77d43a50()
USER32! 77d43b1f()
USER32! 77d43d79()
USER32! 77d43ddf()
nsAppShellService::Run(nsAppShellService * const 0x00a59db8) line 524
main1(int 1, char * * 0x002e2638, nsISupports * 0x0099f2f0) line 1303 + 32 bytes
main(int 1, char * * 0x002e2638) line 1716 + 37 bytes
mainCRTStartup() line 338 + 17 bytes
KERNEL32! 77e814c7()

Comment 16

[Firefox Bug Husbandry Bot]

https://bugzilla.mozilla.org/show_bug.cgi?id=1472046

Move all DOM bugs that haven’t been updated in more than 3 years and has no one currently assigned to P5.

If you have questions, please contact :mdaly.