Re: [hatari-devel] Incomplete emulation of IKBD $0A command

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 30/10/2021 à 13:24, Thorsten Otto a écrit :
> Hi,
>
> The IKBD command $0A (causing the IKBD to report cursor keys instead of 
> mouse movements) takes two parameters for the thresholds. These are not 
> considered by Hatari, instead it behaves as using threshold values of 1.
>
> This can be seen eg. with gulam, where every mouse movements scrolls the 
> history.
>
> BTW, does anybody know what happens if those thresholds are set to zero?

Hi

from an annotated dump of the IKBD rom I made in 2012, I see this code :

F5D2 : 91 C5            "  "            cmpa    X00C5
F5D4 : 25 03            "% "            bcs     LF5D9
F5D6 : 72 01 C6         "r  "           oim     #$01X00C6
F5D9                            LF5D9:
F5D9 : 39               "9"             rts

where RAM 0xC5 will contain deltax/deltay value and reg A will contain 
x/y distance.

this would translate to :


if ( x_distance >= deltax )
{
  or #1,$c6
}


So I would say that delta_x =0 or =1 should have the same effect when 
x_distance is >0, it will report a key left/right for every change of 1 
unit in x distance.


This matches the code in IKBD_SendCursorMousePacket(), except 
KeyCodeDeltaX/Y is not taken into account, it just does for example :

if (KeyboardProcessor.Mouse.DeltaX>0)
but correct would be
if (KeyboardProcessor.Mouse.DeltaX>KeyCodeDeltaX)

I think I can provide a patch for this, but I don't know the "gulam" 
program ; where can I get it to test the patch ?

Nicolas