That may not answer your question, but the LEDs on an AT keyboard can be software-controlled.

One of the terminal programs, Telex or Telemate, did feature this in the form of a running light..

Jo22 wrote on 2022-11-22, 08:08:

That may not answer your question, but the LEDs on an AT keyboard can be software-controlled.

One of the terminal programs, Telex or Telemate, did feature this in the form of a running light..

The how does the keyboard know if it's supposed to add fake shift up and downs for numlocked keys?

superfury wrote on 2022-11-22, 00:29:

And an AT keyboard image for reference:
https://en.wikipedia.org/wiki/Model_M_keyboar … IBM_Model_M.png

That is not the original AT keyboard layout. This is:
https://en.wikipedia.org/wiki/Model_F_keyboar … inal_layout.png
As you can see, there's no separate navigational block, only the numpad.

Now if you re-read the "Fake shifts" section keeping in mind that the "old AT keyboard" they speak of is the Model F keyboard, while the Model M is "the new thing", it kinda start making sense.

The how does the keyboard know if it's supposed to add fake shift up and downs for numlocked keys?

The keyboard knows if its LEDs are lit. But their state is not toggled by the keys themselves, instead the key sends its scancode to the computer and the OS responds "turn this LED on/off".

Azarien wrote on 2022-11-22, 15:48:

That is not the original AT keyboard layout. This is: https://en.wikipedia.org/wiki/Model_F_keyboar … inal_layout.png As you can […]
Show full quote

superfury wrote on 2022-11-22, 00:29:

And an AT keyboard image for reference:
https://en.wikipedia.org/wiki/Model_M_keyboar … IBM_Model_M.png

That is not the original AT keyboard layout. This is:
https://en.wikipedia.org/wiki/Model_F_keyboar … inal_layout.png
As you can see, there's no separate navigational block, only the numpad.

Now if you re-read the "Fake shifts" section keeping in mind that the "old AT keyboard" they speak of is the Model F keyboard, while the Model M is "the new thing", it kinda start making sense.

The how does the keyboard know if it's supposed to add fake shift up and downs for numlocked keys?

The keyboard knows if its LEDs are lit. But their state is not toggled by the keys themselves, instead the key sends its scancode to the computer and the OS responds "turn this LED on/off".

OK. So it looks at the currently set keyboard NumLock LED and when it's set it adds the fake shifts when:
- Release left/right fake shift when off w/ left/right shift(temporarily undoing the shift).
- Press left fake shift when on w/o shift(forcing the shift).
- NumLock pad slash: release left/right shift w/ left/right shift(temporarily undoing the shift).

That's what my keyboard emulation is doing right now.

Is that correct?

Hmmm... I've read here that the Num Lock and Scroll Lock keys depress the control key.
Is that true? I've just implemented a plain LCTRL and RCTRL release when pressing the button, pressing it again when the key is released(only as long as rctrl and lctrl is still pressed).

Is that correct behaviour? I can't seem to find out anything about it in the other sources?

I don't know all the details. However, this paper by Microsoft gives explicit scancode sequences for at least some of those special cases:

https://download.microsoft.com/download/1/6/1 … 6c/scancode.doc

No special treatment given to Scroll Lock, which probably means it's not required.