BBB (16:57:45)

arrakis (16:50:04)
NOW BOOKING FOR REVISON

Agemixer (14:49:05)
when switching proc's i mean

Agemixer (14:48:08)
timers and ROM mode set by the actual program. So the less the scheduler touches on things for the "debug process" the better anyway

Agemixer (14:44:33)
then this makes CIA and other timings registers pretty much untouched, so no matter which basic program you run!! Could be rare exceptions, timers for some ML code executed with sys command can break that "sandbox" of the same

Agemixer (14:39:50)
A little bit patched version (ROM copied to RAM, then patched) might be needed, the scheduler gets active then somewhere between basic commands execution

Agemixer (14:37:07)
and in the end result it turns out a very simple memory swapper prog, most of the time

Agemixer (14:34:14)
just single colored, no voices, no i/o memory at all, limited memory locations to swap, then it runs fast

Agemixer (14:32:14)
but to make the scheduler work properly, it swaps so many bytes in memory, it will be sluggish, and quite inefficient even. The less i/o or vars area is required in "code and debug" process side, the less swapping is needed, the better efficiency.

Agemixer (14:25:31)
while the another "process" is executing the actual basic program..