* new WORKING machines ---- Magic Purple [David Haywood, Jose IBI, Arcade Vintage, Recreativas.org, DU]
* sound balance seems off (nw)
* use XTAL (nw)
* simplify things a bit (nw)
* no 'language' dip on this (nw)
* not related (nw)
* Removed BAD_DUMP flag from josvolly MCU dump - there's no reason to believe the MCUs contain different programs, and Allumer were known to use different silkscreen on identical parts (would still be nice to get confirmation).
* Connected gsword MCUs and correted connections for DIP switches, player inputs, start buttons, coin chutes, and coin counters.
* Lots of guesses due to lack of measurements, schematics, or even good PCB pictures.
* Demoted gsword to MACHINE_NOT_WORKING and MACHINE_IMPERFECT_SOUND as there are issues with credits registering and sound.
* Added extensive notes.
(nw) I'm pushing this in case someone else wants to take a look. The
actual MCU emulation seems to be working, and the MCUs are communicating
without errors. The MCUs checksum three quarters of their program ROM
(passing) and the other quarter all looks good. The problems may stem
from some status flag not propagating back to the main CPU, or possibly
something in the interrupt logic that was working before because serial
communication delays were inaccurate. Although previously a set of
four MCUs were simulated, there are only three actually present. The
second I/O MCU was mapped at a location never accessed by the sub CPU.
* Correct palette problems and silence log output
Monitor autodetection avoids false colors [on a real machine, this would require rewiring].
Corrected palette problems (non-working flag). Also, video levels from DEC's documentation turned out too dark. Converted printf to logerror.
* Old_palette removed
* More fixes that slipped through (NW)
* Update rainbow.cpp
* Update rainbow.cpp
* seta2: document raster effect findings so far, it's about as illogical as the rest (nw)
* improvements, some stuff still broken by yofset change tho (nw)
* fixup problem in previous (nw)
* preliminary raster type stuff (nw)
* refactor and fix silly mistake (nw)
* not really happy with this, but not much evidence to go on at the moment (nw)
