🙏 Testers wanted 🙏

[hrvach]

I am working on extending support for some devices that had issues historically and would kindly ask for some help with testing the upcoming releases.

Even if your keyboard/mouse works perfectly, I'd still appreciate your input on an upcoming release to make sure nothing got accidentally broken for you.

If your keyboard/mouse wasn't working, your input is especially appreciated - despite the efforts, it's hard to fix issues without access to the hardware that doesn't work. Your help is much appreciated!

[Update] Newest version:

deskhop.zip

Thank you!

[p3lim]

No dice with my keyboard.

[chri2]

First item to test - testing-release.zip

I'll try as soon as I'll find time. But: could you upload the source as a branch so it'd be possible to compile it with my own settings? With the default settings the mouse for sure is not usable in my setup and testing is quite difficult.

[hrvach]

I'll try as soon as I'll find time. But: could you upload the source as a branch so it'd be possible to compile it with my own settings? With the default settings the mouse for sure is not usable in my setup and testing is quite difficult.

Of course I'll upload the source, just have to tidy it up a bit so it's not totally embarrassing :)

[raldnor]

Hi Hrvach,

I just tried this release and the rotary (volume) encoder works, including mute (a press on the encoder), but fn+esc does not work with this firmware. I mapped fn+esc to the power button in the QMK firmware.

Cheers!

[hrvach]

I just tried this release and the rotary (volume) encoder works, including mute (a press on the encoder), but fn+esc does not work with this firmware. I mapped fn+esc to the power button in the QMK firmware.

That's kind of expected, since system control section is next on my to-do list - I'll update here when I get it done. Thanks!

[jalmeroth]

Hi @hrvach ,

for me the device is constantly resetting or switching the active board back and forth.

This is what I saw when pressing an 'a' on the keyboard.

Here's the full pcap (Wireshark) file for this one session:
deskhop_testing_release.zip

Hope this helps,
Jan

[hrvach]

@jalmeroth apologies, I'm a dumbass and mistakenly uploaded debug version of the firmware. Updated, sorry! :)

[hrvach]

It's stable w/o adapter, even when the adapter gets connected, but once I hit a key it starts to go wild

I updated tinyusb in this release, maybe something changed? Or I just made a bug I can't find :)
testing-release-3.zip
This one is built with the previous (current prod) version of tinyusb. Same thing?

[hrvach]

It's stable w/o adapter, even when the adapter gets connected, but once I hit a key it starts to go wild

Also can you please describe your keyboard/receiver usb vid/pid? I need to get one :)

[jalmeroth]

Also can you please describe your keyboard/receiver usb vid/pid? I need to get one :)

All of this info has been collected here: #47.
I will test the new release later and report back

[hrvach]

Also can you please describe your keyboard/receiver usb vid/pid? I need to get one :)

All of this info has been collected here: #47. I will test the new release later and report back

Thank you so much! I've found at least one bug in the meantime (despite it working for me with no issues), please hold on while I fix it :)

[jalmeroth]

v4 is stable for me now 👍 But the parsed/sent keyboard values are not sensible yet.

[hrvach]

This one should be better.

testing-release-4.zip

[chris-010]

no problem so far

[Faisalalali]

Hello, new tester here.
I'm noticing that mouse movement is inconsistent on both Windows and Linux. Kinda like if the acceleration is skewed to the bottom right somehow. I tested a typical mouse and a Logetich MX mouse and they both show same behavior.
Has the mouse movements emulation changed in this version?

[hrvach]

Hello, new tester here. I'm noticing that mouse movement is inconsistent on both Windows and Linux. Kinda like if the acceleration is skewed to the bottom right somehow. I tested a typical mouse and a Logetich MX mouse and they both show same behavior. Has the mouse movements emulation changed in this version?

Well caught - respect, it's a regression of a previously fixed bug. And embarrassing enough, I've been using it without noticing :)

[hrvach]

New version with LOTS of changes is up for testing while I tidy and commit the code. Most keyboard shortcuts related to configuration have been removed in favor of a web GUI.

Changes and updates:

1. UART reworked to use DMA without the need for CPU to worry about serial comms.

2. Single unified firmware, just one .uf2 which should hopefully autodetect where it's running.

3. One box should be able to update the other if there is a higher version detected running. You will need to only update one of. them, the other will copy its firmware and update itself.

4. Firmware upgrade can now be done easier, with board presenting itself as a usb stick while still running, so you can use mouse/keyboard and not have to plug them directly - everyone hates tracking cables.

5. There is a new GUI "configuration mode" activated by "left shift + right shift + C + O". This causes the box on the kbd side to reboot and present itself with additional configuration HID endpoint and a mass storage device (like a USB drive). On that drive, you'll find a "config.htm" file you can open using Chromium/Chrome - sadly Firefox has no webhid :-(( This presents a web form to configure the device in an easy way, without having to remember 400 weird shortcuts. Hit connect first, device should be listed (but only if it's in configuration mode).

6. Linker script changed, build will have a new script that calculates crc32 and updates the binary so the box can verify flashing process validity.

7. HID parser reworked, some nkro keyboards should be supported.

I probably forgot something. Thanks for any feedback and testing...

[hrvach]

it's not worse than the stable release?

Hm, I used it all day and the only thing somehow worth is that I can't adjust the Speed parameters finer. But beside that it works.

Ah, maybe one more thing I didn't look into: I could'nt get used to the INS and POS1 key position on my actual keyboard and used udev hwdb settings to change there meaning. This doesn't work anymore. I'll look into that later, but I guess it'll just be changed USB vendor:product numbers...

The speed I'll fix, thanks for noticing! The vid:pid are different only while in configure mode, which is not supposed to be left running (for security reasons). I'll consider adding a timeout and a flashing light or something... :)

[chri2]

The vid:pid are different only while in configure mode, which is not supposed to be left running (for security reasons). I'll consider adding a timeout and a flashing light or something... :)

That's a good idea. I just closed the browser and that's why it still had been on. I found the exit button now and the keyboard keys are back at their designated places ;-) - thanks a lot for the explanation!

[hrvach]

The vid:pid are different only while in configure mode, which is not supposed to be left running (for security reasons). I'll consider adding a timeout and a flashing light or something... :)

That's a good idea. I just closed the browser and that's why it still had been on. I found the exit button now and the keyboard keys are back at their designated places ;-) - thanks a lot for the explanation!

Thank you, you've managed to suggest 5 things I missed and 3 things to improve. You're the new head of quality assurance! :)

[chri2]

Haha, you know the saying "Die Geister, die ich rief…"

Your project is amazing and I follow it with great joy! Thanks a lot!

BTW: A question about security came to my mind: I really liked the idea that the device could not be compromised without the need of touching the hardware (well, yes, there's already been that hotkey to switch to upgrade mode…). The setup becomes by the automatic update between the two picos less secure.

For changes to the software or configuration I'd love to have a security mode that needs the user to (there do not seem to be many options) push the button to confirm actions (or if there are options, short pins, touch something - whatever makes it necessary to actually have access to the hardware to confirm actions).

A nice example for this kind of added security is the Nitrokey3: it blinks and waits for a touch confirmation if configured to do so.

[chris-010]

Here are my findings:

Windows 10
Chrome Version 126.0.6478.62

• installed test fw on both nodes
• changed to config mode when hotkey pressed
• Problem1: CAPSLOCK stays active, can just be disabled by pressing capslock on keyboard connected to deskhop
• Problem2: After disabling capslock, new strage behavior, it seems that shift is still active. You can't see this on keyboard input, but on mouse input, cause text is getting marked from outgoing from the coordinates where Problem1 appears.
• Workaround: CTRL+ALT+DEL -> ESC
• opened config.htm on a Windows 10 with Chrome
• so far so good, but not able to connect