The new Microsoftslop copilot key always sends the following key-sequence when pressed:

copilot key down: left-shift-down left-meta-down f23-down f23-up left-meta-up left-shift-up
copilot key up: <null>

This means there’s no real key-up event when you release the key --> it can’t be used (properly) as a modifier like ctrl or alt.

The workaround is to send a pretend key-up event after a time delay, but then you mustn’t be too slow / fast when pressing a shortcut.

tldr: AI took a perfectly working modifier key from you.

— edit —
Some keyboards apparently do the “right” thing and don’t send the whole sequence at once, you can remap those properly with keyd, see: https://github.com/rvaiya/keyd/issues/1025#issuecomment-2971556563 / https://github.com/rvaiya/keyd/issues/825

copilot key down: left-shift-down left-meta-down f23-down
copilot key up: f23-up left-meta-up left-shift-up

this will still break left-shift + remapped copilot and left-meta + remapped copilot, but RCtrl remaps should work as expected

    • brsrklf@jlai.lu
      link
      fedilink
      English
      arrow-up
      16
      ·
      edit-2
      7 hours ago

      So to be clear, this key sequence is just how windows interpret the key, the hardware is exactly the same and any other OS can still use it as the context menu key?

      Edit : oh, just saw the thing about the linux workaround. So no, they actually fucked it up on hardware level. Wow.

      • attero@discuss.tchncs.deOP
        link
        fedilink
        English
        arrow-up
        16
        ·
        7 hours ago

        nope, the hardware / keyboard controller sends a complete key sequence instead of a distinguishable key-up and key-down event. The OS can interpret that sequence as it sees fit, but you loose the physical key-up signal when you release the key with your finger.

        • Corngood@lemmy.ml
          link
          fedilink
          English
          arrow-up
          5
          ·
          3 hours ago

          That’s insane. Even if they did this intentionally to be as difficult as possible, they locked themselves out of being able to detect long presses?

          • attero@discuss.tchncs.deOP
            link
            fedilink
            English
            arrow-up
            3
            ·
            edit-2
            3 hours ago

            I’ve made an update edit: Some hardware vendors fucked up when to send the key-up-sequence apparently so now every keyboard can behave differently. I don’t know if this makes the situation better or worse.

    • MagicShel@lemmy.zip
      link
      fedilink
      English
      arrow-up
      3
      arrow-down
      1
      ·
      edit-2
      7 hours ago

      AI PC / Copilot+ label

      Okay that sounds solvable, at least. I mean, I hate it, but it seems that a person is getting what they pay for here. Thanks for the heads up. Hopefully there will be plenty of non-AI PC / Copilot+ computers and keyboards.