hasu@tmk

It seems to be reasonable and third soution would be preferable. https://github.com/qmk/qmk_firmware/issues/941

This problem is not observed anymore or seldom now. At least on Linux.

What's your gcc version? What if checkout 27e6c8d(which is used with online editor) and buid firmware? I use avr-gcc 5.4.0 on ubuntu 20.04 to build firmware for reference. ``` $...

which is gcc version of manjaro Linux? You see any warning or error during build process? I'll try new version of GCC when I have a chance.

I will have to check GCC release notes about AVR closely. Something changed between 5.4 and 9.3 and it exposes bug or incompatible of my code probably. https://gcc.gnu.org/releases.html

I confirmed the problem on Archlinux with avr-gcc 12.1.0. At this time workaround seems to be removing `-ffunction-sections` option. I'm not sure how this option causes inside. ``` diff --git...

Currently GCC 7.3.0 supported officially by Microchip is recommended. Confirmed that 5.4.0 also works. https://github.com/tmk/tmk_keyboard/wiki#avr-gcc # GCC 5 and 7 Confirmed that GCC 5.4.0 and 7.3.0 works adding this change...

what is actually you hardware? Genuine GH60 pcb or anything? which pins did you try as 'other pins'?

Note that pin names on Pro Micro is different from ones in firmware codes. Yes, D4 pin on Pro Micro **happen to be** identical to PD4(or D4) in firmware. But...

Tried your hex file and got same result, '6' is registered on Linux and Windows as expected. This is my hex file and almost same. I think you built firmware...