redox_w firmware broke with switch to UART

[achatham]

I could no longer produce a working firmware via configurator, while it worked before. I've built maually and verified that it broke between 0.15.25 (works) and 0.16 (does not work), and head also doesn't work. I've tried git bisect but must not be doing it right yet. I'll keep searching to try to figure out the cause.

The bad firmware flashes successfully but no keypresses register.

[achatham]

Confirmed it's from c122688, which changed UART. I'm on windows 11 and the receiver being flashed has a USB-C connection.

[achatham]

Based on the variables, I'm guessing this is about USB-C and not redox_w. But I'm at my limit of being able to debug the next step, as I'm not too familiar with serial comms.

[drashna]

@fauxpark any insight?

[tomaszj]

I can confirm this — compiling firmware on 0.15.25 works as expected. I use a Redox Wireless made by Falbatech with a receiver wtith a USB-C socket.

0.16.9 tag and master branch locally are broken, whether compiled on a MacBook Air or via Docker — I have tried both routes.

Using firmware from QMK Configurator suffers from the exact same issue.

Note that I can get keys fine after erasing EEPROM and then flashing. Keys stop being sent after a single reset.

cc @mattdibi

[danielmitchell]

I'm also experiencing this issue and have been unable to update my firmware using QMK Configurator since the issue was introduced. Thanks for addressing the issue since my layout is causing RSI and needs to be updated. Hoping this can be fixed soon.

[achatham]

I'm curious what the common thread is. Do others experiencing this have a usb-c connection? A redox-w?

[danielmitchell]

Yes, I'm using a usb-c connection with a redox-w

[mffseal]

I have the same problem. The firmware published by the original author of the keyboard works fine.
Firmware downloaded from the configurator will break after the receiver restarts.

[tomaszj]

The fix is now in master @achatham @danielmitchell @mffseal