(solved) Re: mouse scrolling activates back/forward in browser

[Ronald Klop]

Van: Michael Gmelin <freebsd at grem.de>
Datum: zondag, 18 april 2021 22:49
Aan: Ronald Klop <ronald-lists at klop.ws>
CC: Michael Gmelin <freebsd at grem.de>, freebsd-x11 at freebsd.org
Onderwerp: Re: (solved) Re: mouse scrolling activates back/forward in browser
> 
> 
> 
> On Fri, 16 Apr 2021 21:38:29 +0200
> Ronald Klop <ronald-lists at klop.ws> wrote:
> 
> > On 4/16/21 9:14 PM, Michael Gmelin wrote:
> > >
> > >   
> > >> On 16. Apr 2021, at 21:05, Ronald Klop <ronald-lists at klop.ws>
> > >> wrote:
> > >>
> > >> On 4/15/21 9:11 PM, Michael Gmelin wrote:  
> > >>>>> On 15. Apr 2021, at 20:54, Ronald Klop <ronald-lists at klop.ws>
> > >>>>> wrote:  
> > >>>>
> > >>>> On 4/15/21 9:53 AM, Ronald Klop wrote:  
> > >>>>> Hi,
> > >>>>> When I use the scrollwheel of the mouse my browser triggers
> > >>>>> back/forward. What happens is this: If I scroll up (button 4)
> > >>>>> after scrolling down (button 5) it triggers button 8 and 9. I
> > >>>>> see it generating a ButtonPress 8/9 if I go down and than up.
> > >>>>> And a ButtonRelease 8/9 if I first go up and than down. I have
> > >>>>> no config in /etc/X11. Xorg logs:
> > >>>>> https://www.klop.ws/Xorg.0.log This happens on the trackpad of
> > >>>>> the laptop and with a bluetooth connected mouse. What makes it
> > >>>>> a little more complex is that this is running FreeBSD 14 in
> > >>>>> VirtualBox on Windows 10 with. NB: It happened in Virtualbox 5
> > >>>>> and now in 6 also. This started around a year ago.  
> > >>> That’s because we switched to udev/libinput by default around
> > >>> that time.  
> > >>>>> https://lists.freebsd.org/pipermail/freebsd-ports/2020-March/117879.html
> > >>>>> But didn't have time to look into it. Any thoughts or similar
> > >>>>> experiences? Regards,
> > >>>>> Ronald.  
> > >>>>
> > >>>>
> > >>>> Some more googling found this:
> > >>>> https://forums.freebsd.org/threads/mouse-wheel-strange-behavior-with-firefox.57494/
> > >>>>
> > >>>> Just removed button 8 and 9.
> > >>>> xmodmap -e "pointer = 1 2 3 4 5 6 7 0 0 10 11 12"
> 
> Hi Ronald,
> 
> I could reproduce the issue in VirtualBox. When checking the output of
> `libinput debug-events' I can see button press and release events for
> BTN_SIDE and BTN_EXTRA (which maps to back and forward by default) every
> time scrolling changes direction. After that, the proper AXIS events are
> shown. Scrolling left and right also emits those events - therefore X11
> sees button 8 and 9 instead of buttons 6 and 7. So far so good. The
> problem also happens when an evdev driver is used, so it's not a
> libinput problem.
> 
> Instead of removing buttons with xmodmap, you can also use
> 
>   xinput set-button-map "IntelliMouse Explorer" 1 2 3 4 5 6 7 0 0
> 
> e.g., in ~/.xinitrc, the effect is the same.
> 
> An alternative approach is to change virtualbox to use the USB Tablet
> driver:
> 
>   VBoxManage modifyvm "VM Name" --mouse usbtablet
> 
> (or using the VirtualBox GUI) and then, inside the VM, do:
> 
>   pkg install utouch-kmod xf86-input-evdev
>   echo "utouch_load=YES" >>/boot/loader.conf
> 
> Create a file called /usr/local/etc/X11/xorg.d/99-vbox.conf,
> containing:
> 
>   Section "InputClass"
>       Identifier "VirtualBox Mouse"
>       MatchIsTouchscreen "on"
>       MatchDevicePath "/dev/input/event*"
>       Driver "evdev"
>   EndSection
> 
> and reboot, then start X.
> (in case you're not rebooting, do "kldload utouch; kldunload uhid").
> 
> You also need to make sure that Mouse Integration is enabled in
> VirtualBox.
> 
> This will essentially have the same net-result: Vertical scrolling
> works okay, no forward/back button issues, no horizontal scrolling. So
> it's not necessarily worth the trouble, but I wanted to document it
> anyway for future reference.
> 
> Ideally, there would be a solution that would allow the mouse to
> actually emit horizontal scroll events. I suspect that the reason why
> this isn't working with PS/2 emulation is in our IntelliMouse Explorer
> driver, but I don't know for sure.
> 
> Best,
> Michael
> 
> -- 
> Michael Gmelin
> 
> 
> 


Hi,

Thank you for your effort in reproducing the issue and the excellent explanation. I will try the utouch options in the near future and see what works best for me.

Regards,
Ronald.