30167 - menus should release grab immediately after click

Status: NEW

(Core :: XUL, defect, P3)

Description

[Akkana Peck]

I can no longer set a breakpoint in gdb for anything called from a menu, because
the menu does a grab and doesn't release it immediately, so when I hit my
breakpoint, the grab is still in effect and I can't get focus to the gdb window.
I have to rlogin from somewhere else and kill gdb.  For example, run the editor
on a file, set a breakpoint in nsHTMLContentSinkStream::Write, and do
Debug->OutputHTML.

The grab should be released immediately after I click on the menu, before the
menu's action is called.

This is fairly new; I used to set breakpoints like this in gdb all the time.

Comment 1

[Mike Pinkerton (not reading bugmail)]

want to try to fix for akk, but will need some help here. hyatt?

Comment 2

[Mike Pinkerton (not reading bugmail)]

hyatt says that the mouse grabbing is done not in XPMenu code, but in the gtk 
widget code. pushing over to pav who probably is the one who regressed it.

Comment 3

[Stuart Parmenter]

we call the grabbing code from the xpmenu code.  Due to the nature of the async
reflow code, this is a hard problem to solve.  You have to get the timing just
right.  I would recommend doing something like what I have here:



akkana -- for debugging purposes, if you want to comment out the calls to:

gdk_pointer_grab
gdk_keyboard_grab
gdk_pointer_ungrab
gdk_keyboard_ungrab

in your tree in widget/src/mozilla/gtk/nsWindow.cpp

then we won't grab your mouse and keyboard

Comment 4

[Peter Trudelle]

Mass moving M17 bugs to M18

Comment 5

[Peter Trudelle]

mass-moving all bugs to m21 that are not dofood+, or nsbeta2+

Comment 6

[Mike Pinkerton (not reading bugmail)]

does this have any end-user effects? pushing out.

Comment 7

[Dan Mosedale (:dmosedale, :dmose)]

This bug screws me regularly when I'm trying to test something from a bookmark. 
If I'm on my laptop, which doesn't have open network ports, I often have to
completely kill my X server to recover.

Comment 8

[Robert Braddock]

Yes, this has user effects. Moz takes about five minutes to close a window or
add a a bookmark (my bookmark has a couple copies of an old disorganized
bookmarks list and is very long), and blocking access to X for that long is
ridiculous. Regardless, it seem to me that if you're going to grab mouse and
keyboard, you have something of an obligation not to do it longer than
absolutely necessary.

Comment 9

[Dan Mosedale (:dmosedale, :dmose)]

Another end-user effect: if I've got a menu dropped down, and I want to use one
of the window manager shortcut keys to move to another desktop to check on
somethimg, I can't do it.

Comment 10

[Christian :Biesinger (don't email me, ping me on IRC)]

*** Bug 118866 has been marked as a duplicate of this bug. ***

Comment 11

[mozilla]

Another end-user effect: it's damned annoying.  If I have 1 menu dropped, and I
want to look at another menu or button, I have to click TWICE.

Always makes ppl not used to it wonder why the screen isn't moving, then 5min
later they realise the button they clicked hadn't been clicked yet.

BTW, I'm on WinXP.  It's not just Linux

Look, this is 3yrs old in a few months.  What's with this?  There are very few
programs around nowadays with user annoyances like these.  These should have a
higher priority than things like giving html forms the XP Visual Style.  Can't
some of the paid netscape workers be forced to spend a week clearing out all
these small things that make people want to switch to another browser?  Grandmas
don't say 'IE has security holes'.  They say 'I have to click twice for Mozilla
to listen to me'

Comment 12

[Ian Nartowicz]

This may well be fixed, or at least partially fixed.  See bug 66834 and bug
187575.  Certainly comment 11 is being addressed there.

Comment 13

[Sylvestre Ledru [:Sylvestre]]

Moving to p3 because no activity for at least 1 year(s).
See https://github.com/mozilla/bug-handling/blob/master/policy/triage-bugzilla.md#how-do-you-triage for more information

Comment 14

[Stephen Donner [:stephend] Not actively reading bugmail]

Mass-removing myself from cc; search for 12b9dfe4-ece3-40dc-8d23-60e179f64ac1 or any reasonable part thereof, to mass-delete these notifications (and sorry!)

Comment 15

[BugBot [:suhaib / :marco/ :calixte]]

The bug assignee didn't login in Bugzilla in the last 7 months.
:enndeakin, could you have a look please?
For more information, please visit auto_nag documentation.

Comment 16

[Neil Deakin]

Don't know if this is still an issue, but wouldn't have any enduser impact.