[Coco] The Hobbit ported to 6809 (Dragon64 and coCo2 64k)

[Pere]

Hi Art,
exactly, I have tested the Hobbit on VCC without issuing the POKE&HFFDE,0 
and it works
fine too.
The program first switches to map1 and copies Basic and cartridge, if not 
already in it.
The data from structures in the range $e000-$fee3 is well updated in CoCo3 
despite
the poke&Hffde,0
I will compare ($ffff) against $B4 (the Dragon signature) to detect if a 
CoCo, any kind,
is running the program.
Things are getting easier!

thanks a lot
cheers
pere

-----Mensaje 
original--------------------------------------------------------------------------------
Arthur Flexser flexser at fiu.edu
Sun Aug 21 19:21:15 EDT 2016
The check at $FFFF should work fine.
I would assume that the program copies ROM to RAM if it needs to store data
in high memory, or at least puts the CoCo in RAM mode.  That should mean
that the poke to $FFDE is unnecessary.  And I think that having a separate
test for the CoCo 3 will also turn out to be unnecessary, now that you've
fixed the problem with the interrupt vectors being overwritten.  That is,
the program shouldn't care if it finds the CoCo 1/2 value or the CoCo 3 one
in $FFFF.
Art