search

zmkfirmware / zmk

ZMK Firmware Repository
https://zmk.dev/
MIT License
2.67k stars 2.74k forks source link

`&kp LPAR` generates [`LSHFT` Press, `(` Press, `LSHFT` Release, `9` Release] keycode sequence #2344

Closed glenn20 closed 3 months ago

glenn20 commented 3 months ago

The Problem

Pressing a key mapped (&kp LPAR) to the LPAR symbol (or any other symbol key which is reached via SHIFT on a conventional keyboard) generates the following key sequence:

(tested using xev application on debian linux).

Expected keycode sequence:

Impact

This leads some applications on Linux (esp. terminal applications such as Gnome Terminal and Tilix) to sometimes register the key as a 9. The misidentification does seem to be timing dependent (andf may be affected by terminal settings). Eg. the problem occurs more frequently for "shifted symbols" assigned to home row mod keys on my SYMBOLS layer.

If the output key sequence is by design (intention):

If this is not the intended behaviour, any advice on addressing this through my configuration would be very welcome.

Keyboard: Ferris Sweep (nice_nano_v2 cradio_left/right) ZMK version: Tested with main branch latest commit 7be955ff (and over the last few months) Computer: Debian linux 12 (bookworm) (connection via bluetooth) Config: https://github.com/glenn20/zmk-sweep/blob/main/config/cradio.keymap

caksoylar commented 3 months ago

See #759 and #2334.

glenn20 commented 3 months ago

@caksoylar Thanks for the links. It was really hard to find the right words to search in issues. Very encouraging to see PR2334 in the works. I'll mark this closed.

glenn20 commented 3 months ago

Ah... as it turns out, the problem I identify above is not causally linked to the impact I was seeing. After applying #2334, the correct sequence of LSHFT Press, ( Press, ( Release, LSHFT Release is generated (as reported by xev), but the terminal programs continue to produce the unshifted key (9). Some of the remedies described in #759 may be the best workaround for now. Thanks @caksoylar .

glenn20 commented 3 months ago

On further investigation... #2334 does substantially improve the issue for my non-home row keys. Those on home-row mods are still susceptible to misidentification by applications (eg. gnome-terminal).