6300 - [PP] apprunner: Going thru menu items causes Assertion failure...

Status: VERIFIED FIXED

(Core :: XUL, defect, P3)

Description

[Srinivasa A. Shikaripura]

==================================================
Product:     Mozilla, Release M5
BuildID:     1999050422
Environment: Linux Redhat 5.2, 2.0.36 i586 unknown
Xserver:     Exceed 5.1.3.0 on NT 4.0

How to simulate:
1. Run mozilla-apprunner.sh
2. Click on the test.html displayed on the first page
3. Press "Back"
4. Goto "Tasks" menu item. Slowly go to each of the menu item.
5. Goto "Bookmarks" menu item. Again slowly goto each of the menu items.
6. You can see many assertions being printed on the debug output.

===========Debug output===========
Added page
file:////home/sas/download/package/res/samples/BrowserInitPage.html
to the rdf:history datasource
Got a handle to forward menu item
Setting forward menu item enabled
Obtained MenuItem Back
Setting Back menuitem to enabled
Document
file:////home/sas/download/package/res/samples/BrowserInitPage.html
loaded successfully
nsMenu::MenuConstruct called
nsMenu::MenuDestruct called
nsMenu::MenuConstruct called
nsMenu::MenuDestruct called
nsMenu::MenuConstruct called
nsMenu::MenuDestruct called
nsMenu::MenuConstruct called
nsMenu::MenuDestruct called

Gtk-CRITICAL **: file gtkcontainer.c: line 717
(gtk_container_remove): assertion `widget->parent == GTK_WIDGET
(container)' failed.

[*** Above message repeats 20 times...]

Gtk-CRITICAL **: file gtkcontainer.c: line 717
(gtk_container_remove): assertion `widget->parent == GTK_WIDGET
(container)' failed.

nsMenu::MenuConstruct called
nsMenu::MenuConstruct called
menu_item_activate_handler

[*** Above message repeats 25 times...]

Gtk-CRITICAL **: file gtkcontainer.c: line 717
(gtk_container_remove): assertion `widget->parent == GTK_WIDGET
===========

Comment 1

[Srinivasa A. Shikaripura]

Further comment:

After the above incident, the browser stops responding to mouse.
Another issues to be noted alogn with this is that, the network was
bad when this was observed (every 5th ping packet loss).
After a long time it again responded.

I doubt there may be bugs in the recovery from
network packet losses *also*.

Comment 2

[karnaze (gone)]

Reassigning to Hyatt and CCing Ramiro. There is very apprunner specific
widgetry.

Comment 3

[David Hyatt]

Reassigning to saari.

Comment 4

[saari (gone)]

I know about the assertions, but I've never seen the mouse stop responding.

Comment 5

[stuartp]

the code to fix this is partially checked in to the tree.  i'll add myself to
the cc list on this one

Comment 6

[leger]

Updating QA Contact.

Comment 7

[leger]

Moving all Widget Set bugs, past and present, to new HTML Form Controls
component per request from karnaze.  Widget Set component will be retired
shortly.

Comment 8

[stuartp]

moving to XP Toolkit/Widgets since these are not HTML form controls

Comment 9

[saari (gone)]

Don't have time to deal with this for M8, pushing to M9

Comment 10

[Stuart Parmenter]

*** Bug 9278 has been marked as a duplicate of this bug. ***

Comment 11

[saari (gone)]

Ok, I lied... I am looking at this, but I may or may not get permission for M8

Comment 12

[saari (gone)]

XPMenus sort of negate this. Anyway, I still have a fix that I'd like to checkin,
but I'm not doing it today... pushing to M10

Comment 13

[saari (gone)]

Since this is for native GTK menus, and we're using XPMenus now, I'm pushing this
out past the dogfood "beta"

Comment 14

[Paul MacQuiddy]

dividing up phillips bugs, he no longer works here

Comment 15

[saari (gone)]

Native widgets, yeah right

Comment 16

[Mike Pinkerton (not reading bugmail)]

what should i do with this one, chris?

Comment 17

[saari (gone)]

Kill it. If we ever resurrect native menus I'm sure I can fix it again.

Comment 18

[Stuart Parmenter]

this was fixed when we turned on XP menus.

Comment 19

[sairuh (rarely reading bugmail)]

yup, no longer see those assertions.