I have a Planck board which also uses a fork of TMK. It does not exhibit
the discussed problem behaviour.
I checked what key-codes it sends for @:
Shift_L 0x38
KeyDown @ 0x13
KeyUp @ 0x13
Shift_L 0x38
This is exactly what Atreus is sending, as it looks.
On Sat, Mar 19, 2016 at 8:56 AM, Phil Hagelberg <phil@xxxxxxxxxx> wrote:
S R <rustamoff@xxxxxxxxx> writes:
This, by the way, does not happen on using stock keyboard on the same
machine (rMPB).
Right; because the internal keyboard doesn't have any keys which
generate both a modifier and a normal keycode from a single keystroke.
Each frame in the USB protocol can hold up to six normal keycodes plus
modifiers, so when you press the @ key, it sends a frame that has both
shift and 2 together. Since it only happens in virtualized systems, I
have to assume it's a problem with the virtualization layer's IO
handling where a frame that contains both a modifier and a normal key
can get split in two frames. I can't reproduce with Virtualbox or KVM
and Linux guests, but I don't have a Windows VM to try it on right now.
-Phil
