on 7/23/09 6:46 AM, Konrad Kokoszkiewicz wrote:
I don't know anything about this stuff, but MiNT should behave
exactly
as TOS, especially for programs that are not MiNT aware.
This looks like a bug in MiNT...
If someone could explain the intention of this, there could
possibly be a
workaround in XaAES. Until then you can just disable this feature
in MiNT by
#defining MM_INSERT 0 in keyboard.c, as I wrote before. On my MiNT
the
Bconin-returns are ok now.
It is me who basically wrote the keyboard.c module, but it was so
long ago
that I don't remember such details. I mainly used the Atari
Compendium as
a reference, and own experiments to verify, if the reference is
correct.
It is possible that it is written there and I missed failed to verify
that.
Anyways, if you're sure that it will improve the compatibility of the
keyboard behaviour so that it will behave more like TOS, fix that.
It is
the intention of the module to emulate the TOS behaviour as closely
as
possible.
Speaking of keyboard.c
This section:
# if 0
/* bit 2 of conterm variable decides, whether we
* put the shift status to the buffer or not.
*/
if ((*(uchar *)0x0484L & 0x04) == 0)
c0 = 0;
# endif
Shouldn't this be bit 3 ??
--
Lonny Pursell http://www.bright.net/~gfabasic/