Re: [qubes-users] Pointer lock API

2017-12-25 Thread Vít Šesták
I have encountered such issue and I have found a semi-solution. I usually don't 
use mouse (just mousekeys and touchpad), but for games, I do.

There is an exiting input proxy mechanism. It was originally intended for 
sys-usb->dom0, but other combinations can work, too.

As far as I remember, it does not work with the standard X11 instance in AppVM. 
Instead, I had to run a VNC loopback session. This also required the service 
file in target VM to set DISPLAY environment variable.

If someone is interested, I can look for details. Maybe I have something 
already posted here.

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to qubes-users+unsubscr...@googlegroups.com.
To post to this group, send email to qubes-users@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-users/07c36e37-e419-4cb8-b697-51afde670103%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [qubes-users] Pointer lock API

2017-09-24 Thread yuraeitha
On Sunday, September 24, 2017 at 11:15:28 AM UTC, Leo Gaspard wrote:
> On 09/24/2017 04:50 AM, Nagaev Boris wrote:
> > Hey!
> > 
> > I noticed that Quake3 almost works in Qubes 3.2, but has one annoying
> > issue: the pointer is unmanageable. It looks like the pointer has the
> > "memory" about its previous position: e.g. when I move it left, the
> > game continues moving right. I tried `openarena` tool from the Debian
> > template and also web version of Quake on http://www.quakejs.com/ both
> > have the same issue. Performance of both native and web versions is
> > amazing, the only problem is the mouse.
> > 
> > I think it works incorrectly because it tries to lock the pointer
> > (probably to move it to the screen center) but Qubes doesn't allow
> > AppVMs to manage pointer. X11 has function `XWarpPointer` that moves
> > the pointer. Probably the game uses this function, though I don't know
> > how to check this.
> > 
> > Can Qubes add Pointer lock support for AppVM, please?
> 
> Disclaimer: I haven't tried yet to do this.
> 
> In addition to using the tablet input mode, as Alex advises, you may
> have luck plugging in a USB mouse and passing it through to the VM (if
> you run in non-seamless mode).
> 
> HTH,
> Leo

Sounds to me like a good suggestion, and it will more depend on the guest-OS 
whether it supports USB-mouse, making Qubes support irrelevant. 

Having said that, it would be amazing to sometime in the near future have 
touch-screen support for Qubes and AppVM's. 
If non-touch screens don't disappear with Virtual/Augmented reality spreading 
like wildfire in the coming years, then traditional mouse most certainly is 
going to go away eventually.

With all these new hand movement, brain-wave thought control of software, 
gyroscope finger sensor ring and control. It's hard to see it being a good idea 
not to focus development on these new technologies, which might come to market 
very fast, once the big mobile/laptop developers decide to unleash it.

At which point, we'll be sourly left behind this revolution, which has the 
potential to make software use considerably more convenient, fast and smooth.

It's a bit of a problem, few programmers seem to take anything else but 
keyboard/mouse seriously. Though of course, its hard to program for something 
that hasn't reached mainstream yet, or at best only prototypes or exotic 
commercial releases in the wild.

But considering how long it takes to make touch-screen work properly in the 
Linux community, I do fear for the future once new input methods become 
mainstream. 

I love the Linux community. But if there is anything I dislike, it's the 
conservative, reactive, kind of thinking that takes president. There are so few 
visionaries, proactive, kind of thinkers around in the Linux community in 
general. Though to be fair, Qubes is quite revolutionary and visionary, but I'm 
not talking about Qubes, but rather the desktop environments and their lack of 
proper working input support.

It hardly makes sense for Qubes to implement it, if the desktop environments 
don't support it properly to begin with. In a sense, it makes sense Qubes 
developers don't work on something, which is outside their scope. The problem 
here, is those who are supposed to be responsible for it, the desktop 
environment developers, especially the conservative ones who dislike change and 
don't care about users who have different needs than their own. 

Those are the problem, in my opinion. Albeit can't complain if I don't donate 
to them, but I don't want to donate unless they do a good job either, to which 
many desktop developers most certainly don't. It's a bad never ending circle. 

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to qubes-users+unsubscr...@googlegroups.com.
To post to this group, send email to qubes-users@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-users/4d18d712-6890-4335-9f52-40a98fd3ddf6%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [qubes-users] Pointer lock API

2017-09-24 Thread Alex
On 09/24/2017 04:50 AM, Nagaev Boris wrote:
> Hey!
> 
> I noticed that Quake3 almost works in Qubes 3.2, but has one
> annoying issue: the pointer is unmanageable. It looks like the
> pointer has the "memory" about its previous position: e.g. when I
> move it left, the game continues moving right. I tried `openarena`
> tool from the Debian template and also web version of Quake on
> http://www.quakejs.com/ both have the same issue. Performance of both
> native and web versions is amazing, the only problem is the mouse.
> 
> I think it works incorrectly because it tries to lock the pointer 
> (probably to move it to the screen center) but Qubes doesn't allow 
> AppVMs to manage pointer. X11 has function `XWarpPointer` that moves 
> the pointer. Probably the game uses this function, though I don't
> know how to check this.
> 
> Can Qubes add Pointer lock support for AppVM, please?
> 
It may not be a pointer lock issue, but rather a problem with input
type. Did you try starting the VM with a different mouse input type? The
default seems to be Mouse, while Tablet may solve your problems, like it
did (even if not 100% because of inbuilt android pointer acceleration)
for my Android appVM.

The mouse input type setting cannot be configured via standard Qubes
tools, either the qubes.xml or the manager. You can try by creating a
copy of "appvmname.conf" with a name different from the original (or
Qubes will overwrite it) that you can find in dom0 at
/var/lib/qubes/appvms/appvmname/appvmname.conf.

In the copy find the "devices" element and add "" - this will override the default mouse input. Start the VM
with this conf instead of the default one with qvm-start and check.

It would be a great security concern to let single appVMs manage the
pointer, even if in good faith (misclicks can produce disasters) and
without mentioning the big development problem. The mouse "resource"
will have to be "shared" among potentially several
mouse-management-enabled appvms: who shall win if two appvms decide to
lock the mouse in two separate rectangles?

-- 
Alex

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to qubes-users+unsubscr...@googlegroups.com.
To post to this group, send email to qubes-users@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-users/9e149873-19ab-90ac-ddd8-ee4e90418eed%40gmx.com.
For more options, visit https://groups.google.com/d/optout.


signature.asc
Description: OpenPGP digital signature