The code is just for checking if it's possible, not for using. Most of the java code was AI generated, I think we would need an profile option to specify the desired display ID in XtMapper. If you're interested and have the time to properly implement this, I would really appreciate it!

I certainly don't know what I'm doing, so going to dump all the "research" I've done here in case that's any helpful.

How to get info about all the displays (including external ones)

# Just some commands you can use for inspecting, you may remove the grepping and all for verbose info
cmd display get-displays | grep "Display id"

dumpsys SurfaceFlinger display | grep -a 'parent=Display ' | uniq

# Also from srcpy:
scrcpy --list-displays -e # -e for wireless adb

How to test if you can simulate input on an external display

If you have a samsung device with display out, you can just enable dex and find the display ID. For other phones, you will need to enable "force desktop mode" in the developer settings. It also helps if you enable "pointer location" for testing where the input is going.

Once you have a list of display IDs (sometimes there can be extra virtual displays for input handling), you can test like so:

# Replace -d 0 with other IDs that you found, 0 is usually the main built-in display
input -d 0 swipe 100 200 500 200 3000

What's "working"

• Mapping is possible on the external display (was already, just mentioning)
• Mapped keyboard keys are properly injected in display 2 instead of the phone display.
• Locking/toggling mouse crosshair focus, sensitivity/movement may be buggy or not properly tuned, can't confirm due to lack of proper mouse.

What's not working or needs work

• XtMapper virtual Mouse cursor floats over the main display 0 instead of the desired display. But mouse inputs are being injected in the desired display. I don't have a proper mouse, only a trackpad so couldn't test much.