On Fri, 3 Sep 2004, Eckhard Stolberg wrote:
Does this account for (at least some of) the weird behaviour some games
exhibit when you `fry' them? I always thought it had to do with the RIOT
RAM not being reset, but I hadn't even thought about the timers.
I don't think the RIOT timers have anything to do with frying.
I don't know what causes frying, but it must have something to
do with the state of the RIOT RAM and the state of the program
counter in the 6507.
So either the short power cycle is messing up the RAM and keeps
the program counter intact, or it keeps the RAM intact and just
changes the program counter to anything other than the address
at the reset vector.
We could simulate either or both effects in z26, and see whether they
duplicate the frying effect. I wonder if the program counter or RAM
truly gets randomized, or if just one or 2 bits of it change state...
I still dream of adding a frying option to z26... :)
I'm still dreaming of adding many features to z26 as well,
but I guess one of these days one of us will have to wake
up and start some real programming again in oder to make our
dreams come true. ;-)
I had to get away from it for a while, it was becoming an obsession.
The only way I know how to do anything complicated is to do it 24
hours a day (even dream about it), to the exclusion of everything else,
including work...
--
B.
