* Cleaned up large amounts of duplicated code between the MAME and MESS (AES/NeoCD) drivers
* Removed MVS-only ROMs and devices (SM1 ROM, SFIX ROM, upd4990a RTC, etc.) from AES/NeoCD
* Replaced legacy NVRAM in MVS and NeoCD with NVRAM devices
* Correctly hooked up SM1 ROM bankswitching; the BIOS sound test works now
* Corrected VRAM size
* Sorted the Universe BIOS versions and unified them between MAME and MESS
* Fixed kf10thep ROM loading and decryption which was a horrible hack (and not endian safe)
* Replaced svcplus protection hack ROM patch with a less invasive one, based on FBA
* Started marking code and data that should be moved out into devices
* Various other code and comment cleanups
New clones added
----------------
Kizuna Encounter - Super Tag Battle 4 Way Battle Version / Fu'un Super Tag Battle Special Version [Alex Jackson]
Documented phase shifts 24.7 thru 27.7 degrees in 0.5 increments. Enabled 26.2 degrees as default being a logical choice as the system should be initially calibrated for Hue 1x and 15x (F$) to visual match (~25.7 degrees) displaying 'gold'/'goldenrod' via system's manual pot adjustment, while moderate system runtime automatically creates a higher phase shift.
Results collaborate with additional tech docs (I.E. "Stella Programmer's Guide") placing 1x as 'gold', 2x as 'orange' and 15x (F$) as 'light orange' (hue between 1x and 2x). Colors measured from midpoint of brightness scale. [Robert Tuccitto]
<jvilk> DFJustin: the issue is that if you do not add those targets in, then those targets are not defined when you do a CROSS_BUILD
<jvilk> and Make fails
<jvilk> because targets depend on them
<jvilk> the 'correct' fix would be to have the targets that depend on them depend on the correct location of the tools
<jvilk> because I think what was happening is that they depended on the location of the tools assuming that they were built for the target platform
