Linux Mint 22, spacenavd & SpacePilot (HP, 2007)

[DudeMcGee]

First of all, a lot of things work: In FreeCAD's customize menu, Spaceball Motion, I can adjust the sensitivity, can reverse axes etc. NICE! In the adjacent menu, Spaceball Buttons, I could successfully assign functions to the top row of buttons, which resembles function keys 1 through 6. These assignments actually work :) ! NICE!

But what seems to be 100% unimpressed by all of this, is the great display on the mouse itself, and since it is HP-branded, it shows, without intermission, how proud it is of being a HP product: it displays the HP logo, and that's it. A wonderful logo, but useless.

Is there any way to bring the display to life and show anything meaningful, like key assignments?

Thanks for your advice, help and hints! And a big thank you to the developer team of spacenavd who made it possible, that I can use a 3dconnexion mouse at all!

[jtsiomb]

Short version

Indeed, spacenavd doesn't manage that display at the moment at all, but I have written a standalone library and application to do that: https://github.com/jtsiomb/3dxdisp
You could just run that program to upload a static image to it, or write a full libspnav client that queries information from spacenavd periodically, and updates the display using 3dxdisp to show some useful information about its current status.

Long story

When I wrote that library (3dxdisp), it was meant as a test and maybe the basis of adding space-pilot screen support into spacenavd, but after I did that, I couldn't think of anything useful I could really show in it other than a fancy logo or something meaningless like that. That was because of a number of reasons, most but not all of which are still true today:

1. I've never seen the official driver in action at the time, to get ideas :)
2. spacenavd does not support per-application "profiles", to show which one is currently active
3. spacenavd at the time did not have any meaningful driver-configured keybindings, leaving them all to the application and just supplying the button numbers to it, so there was nothing to show in terms of button mappings (this has changed somewhat, there are some useful binds you can do in spacenavd config).
4. none of the spacenavd-app communication protocols provide any way for the application to notify spacenavd of its internal keybindings, or more generally what it would like to show on the display.

Points 2 and 4 are also not just missing features, but differences in design from the official driver. Spacenavd doesn't have the notion of an "active" application. It transmits events to all listeners, and it's up to them to decide if they have meaningful "focus" in whatever environment they're running, and would make sense to respond to that input. So one application can't monopolize the display to begin with, even if there was a mechanism for the application to send back things to be displayed.

Now there are a few potentially useful things I could conceptualize using the display for in spacenavd. There are internal things like sensitivity change, and dominant axis mode toggle, that can be assigned to specific buttons, and there's keyboard key presses that can be bound to buttons. Those can be useful things to display on screen. Other than that, maybe current values of things again like sensitivity, or maybe the number of clients connected and which protocol each one uses could be relatively useful things to display.

There are a couple of not particularly good reasons I never got around to actually add any of these in spacenavd:

• Very few people actually use a SpacePilot, and there's now a new device with a completely different display. The prospect of reverse engineering the new one too is not particularly high in my priorities. That shouldn't be a blocker, it's just slightly demotivating.
• 3dxdisp introduces a new dependency to libhid and potentially libusb, which I haven't decided how I want to integrate to spacenavd. I intended to write a libusb backend for the main functionality of spacenavd, to improve cross-platform compatibility for USB devices to other UNIX systems beyond Linux, and I always bundled those two things together in my mind, but I never got around to doing that either so far.

So that's the full answer. Feel free to comment on any potential uses of the display which I may not have thought of, or to just add your vote behind implementing any or all of the above :)

[DudeMcGee]

Hey, John,

Thanks for all the information you've been giving to me about the issue on Linux Mint and here about the display support and your design ideas. I'll go through all of that thoroughly and process this material. And I believe, I've understood quite a lot already, although from a non-programmer's perspective.

My plan is, to evaluate what the SpacePilot does on a Windows machine with the current 3Dconnexion driver. I'm especially curious to see what the 3Dconnexion configuration program does. Although this might not be an awful lot, since it can, as you describe exactly, only change its behaviour within the OS realm, and nobody is interested in that at all. Eventually, such a configuration program couldn't do anymore but manipulating the stuff written in the spnavrc file, right? And these settings are irrelevant to the application, like FreeCAD e.g.

Two questions came to my mind:

1. Does the latest SpaceMouse Enterprise do anything conceptually different in terms of display and key management than the SpacePilot?

2. What's the meaning of this section in spnavrc:

# List of additional USB devices to use (multiple devices can be listed)
#
# example:
#    device-id = 046d:c625

And in this context I've seen that the SpacePilot is listed twice as a device, bus 003, device 002 and bus 003, device 012, why is that?

$ lsusb
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 002 Device 002: ID 1d5c:5001 Fresco Logic USB3.0 Hub
Bus 002 Device 003: ID 0b95:1790 ASIX Electronics Corp. AX88179 Gigabit Ethernet
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 003 Device 002: ID 046d:c625 Logitech, Inc. 3Dconnexion Space Pilot 3D Mouse
Bus 003 Device 003: ID 06cb:00bd Synaptics, Inc. Prometheus MIS Touch Fingerprint Reader
Bus 003 Device 004: ID 5986:9106 Bison Electronics Inc. Integrated Camera
Bus 003 Device 005: ID 1d5c:5011 Fresco Logic USB2.0 Hub
Bus 003 Device 006: ID 058f:9540 Alcor Micro Corp. AU9540 Smartcard Reader
Bus 003 Device 007: ID 8087:0032 Intel Corp. AX210 Bluetooth
Bus 003 Device 008: ID 1a40:0101 Terminus Technology Inc. Hub
Bus 003 Device 009: ID 0d8c:0016 C-Media Electronics, Inc. M-AUDIO Uber Mic
Bus 003 Device 010: ID 2537:1081 Norelsys NS1081
Bus 003 Device 011: ID 25a4:9311 USB C   USB C Video Adaptor      
Bus 003 Device 012: ID 03f0:094a HP, Inc Optical Mouse [672662-001]
Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub

Thank you very much, have a great day ;)

[jtsiomb]

such a configuration program couldn't do anymore but manipulating the stuff written in the spnavrc file, right? And these settings are irrelevant to the application, like FreeCAD e.g.

Yes, well indirectly... spnavcfg transmits configuration changes to spacenavd, and spacenavd manages the spnavrc file by itself.

And yes, currently the settings are for all programs. App-specific profiles can be introduced at some point, since the latest iteration of the protocol includes the option for the application to identify itself, although I'm not sure if any applications actually use that. Especially since at the moment that application name isn't used for anything.

1. Does the latest SpaceMouse Enterprise do anything conceptually different in terms of display and key management than the SpacePilot?

On the hardware side? Not really, it's just a higher resolution color display. About software differences, what the official driver displays in each screen, I don't know. I've seen the newer device running on windows with the 3d connexion drivers a few times, I think it displays the active application and keybindings for it, but I don't remember details.

2. What's the meaning of this section in spnavrc:

# List of additional USB devices to use (multiple devices can be listed)
#
# example:
#    device-id = 046d:c625

Allows you to instruct spacenavd to try to open and use a specific USB device, even if it wouldn't otherwise. Basically if you have a multi-axis joystick, that doesn't report itself as a 6dof device but as a HID game controller, you could attempt to make use of that.

It's also often useful for homebrew 6dof devices, that again sometimes just use an off-the-shelf joystick USB controller board, and can't report themselves properly as 6dof devices.

And in this context I've seen that the SpacePilot is listed twice as a device, bus 003, device 002 and bus 003, device 012, why is that?

Bus 003 Device 002: ID 046d:c625 Logitech, Inc. 3Dconnexion Space Pilot 3D Mouse
Bus 003 Device 012: ID 03f0:094a HP, Inc Optical Mouse [672662-001]

The second one you mentioned looks like just a mouse. Are you sure it has something to do with the space pilot? it doesn't look like it.

[DudeMcGee]

Thanks! You're right, the HP Optical Mouse is exactly that. I was so focused on HP and the SpacePilot that I forgot that my normal mouse is just this device. Silly me :)

[DudeMcGee]

You were (of course) right with your assumption about my installation of spnavcfg. I've corrected this, found the thread about repairing the library links with ldconfig, and voila, it works!

[DudeMcGee]

I got lib3dxdisp running :)

Could you, please, give a short explanation how I can convert a picture into the format of logo.h?

[jtsiomb]

Could you, please, give a short explanation how I can convert a picture into the format of logo.h?

I forgot the example program was so simplistic. Still it's not too difficult to change the contents of logo.h, you can do it in gimp:

1. start with a 240x64 image you want to display
2. convert it to black and white
3. export as "C source code header" (.h suffix)
4. open the exported file and copy/paste the data from inside the header_data array, to replace the contents of the logo array in logo.h. Or equivalently just replace the logo.h file with the newly exported one, and rename header_data to logo. Optionally delete everything else in the file, though I don't expect it will cause any problems if you don't bother.

[DudeMcGee]

Great, understood, works!
Have a nice day ;)