Fix Rod's colors on IIc

[frankmilliron]

No description provided.

[inexorabletash]

I want to try more careful fixes here.

For Rod's Pattern, if COUT is going to be used then we should either play nice with the 80-col firmware or disable/re-enable the firmware.

For Matrix, that turning DHIRES off at the end "fixes" Rod's Pattern is surprising, since it's very quickly re-enabled by DeskTop. That may indicate that some of the switches need to get hit in a specific order. In any case, if the DAs modify state they should restore it, so it should exit with DHIRES on, not rely on DeskTop to toggle it back on.

I'll see what I can come up with.

[inexorabletash]

I found a bug that would cause IOUDIS to be in an unstable state when testing. That likely invalidates the discoveries here.