On 11/01/2013 14:57, Helmut Karlowski wrote:
What about Vincent's RAW+ONLCR ?
I finally understood that the main issue encountered by Jo Even was caused by the fact that the VDI and the BIOS console share the same input device: the keyboard. This is why XaAES hacks the BIOS console.
The workaround of putting the console in RAW mode removed the bad keyboard interaction, but introduced the stairs effect.
A possible clean solution would be that XaAES opens /dev/xconout2 at startup, so the console I/O will never reach the keyboard/screen. Opening the TosWin2 console would be mandatory to see the console output (or better, auto-open feature). Conceptually, disabling console I/O would make sense inside the VDI when opening the physical workstation, but this is probably not possible due to the various VDI in use.
Finally, I think that the cleanest and more pragmatic approach would be that XaAES put the BIOS console in RAW+ONLCR mode at startup. I'm not sure if FreeMiNT terminal devices allow to set only that ONLCR flag yet (instead of whole COOKED mode), but adding support should be very simple. Probably simple and safe enough to add it to the 1.18 release. I'm going to try that.
-- Vincent Rivière