Starting off with the 8-bit Apple II, some issues were fixed causing joysticks to misbehave. On IIe-class machines the Zip Chip emulation was fixed so when it’s locked, it stays locked. The Apple IIc Plus switched to the new cycle-accurate IWM and Sony drive emulation, allowing writing to 3.5″ disks to work properly, and enabling the use of both 5.25″ and 3.5″ .WOZ images. (Formatting disks on that machine still has issues, but it’s possible to create pre-formatted ProDOS images in MAME’s file manager so that’s not as big of a deal as it used to be). A bug with the IIc Plus’ VBL registers was fixed that prevented Prince of Persia from working properly.
In Apple II clone world, a regression in the China Educational Computers (CECs) was fixed that broke Pinyin input mode on some machines. The Laser 128 and Franklin ACE machines were next to get the new cycle-accurate IWM and disk drives, and they can boot .WOZ images now too. Finally, preliminary support for the function keys on the Franklin ACE 500 and 2000 machines was added. It’s not 100% but the basics are there.
The would-be successor to the Apple II, the Apple ///, got a major fix with floppy handling, so it’s safe to write to floppy images in the emulation again.
Moving on to the 16-bit Apple IIgs, we started off adjusting our approximation of the machine’s infamous “slow cycles” when accessing certain parts of the system for better compatibility, issues with writing to 3.5″ disks were solved, some corner-case behavior used by the IIgs AOL client was fixed, some even more corner-case behavior involving how banks $E0 and $E1 respond to certain soft switches was fixed, and the IIgs’ clock/BRAM emulation was changed to interface to the device we wrote for the Mac emulation, as the IIgs shares that chip with the Mac. This fixed a few bugs and the IIgs gained some features of the chip it had been missing, including the write-protection setting.
And happening right at the end of the year and just missing the cutoff for this year’s last release, the IIgs got in on the ADB revolution happening on the Mac drivers that I’ll talk about shortly. All current IIgs emulators, including MAME, KEGS, and GSPlus/GSPort, have handled keyboard and mouse input by basically jamming it into the system and inexactly simulating some of the behavior of the Mitsubishi microcontroller (a 6502 derivative) that lives in the real IIgs. Starting with our January release (0.252), the IIgs in MAME runs the real code that lives on that microcontroller and interfaces it to our ADB keyboard and mouse emulation. As a result, everything input-related now works exactly like it does on a real IIgs. Hacks that were necessary to make things like the FTA Nucleus demo work properly before are gone, and games like Wolfenstein 3D that reprogram the microcontroller for special handling will just work.
For Apple II expansion cards, the IIe Standard 80 Column Card got correct emulation (there were some nuances to how it worked that were wrong), and we got preliminary support for the Apple 3.5″ Disk Controller Card (aka the SuperDrive card). It doesn’t work 100%, but it works for booting 3.5″ disks on the Apple IIe, which wasn’t possible until now.
In Mac world, lots of things happened. First up, the Mac LC III and LC 520 were split out of the huge mass that mac.cpp had grown into over the years, and built around a new properly device-ified emulation of the Sonora system ASIC. This makes the driver more lightweight, easier to understand, and easier to maintain going forward. Next up for this treatment were the Mac IIvx and IIvi and their VASP ASIC. And finally came the machines built on the V8 ASIC and its many variants: the Mac LC, LC II, Classic II, and Color Classic. In addition, all Macs got support for booting copy-protected originals via the new .MOOF format supported by the Applesauce imager.
Next up, the aforementioned ADB revolution. The earliest ADB Macs used a GI/Microchip PIC microcontroller as an “ADB Modem” to talk to the ADB port. We’ve had a pretty good simulation of what it does for many years now, but finally one of these chips was decapped and the ROM was extracted. MAME was happy to have that and some of those machines now happily run the real PIC microcontroller ROM to perform those functions. Due to timing issues that we’re working on, not all of the machines that could benefit from this currently do, however. Meanwhile, we’d had good success for a while running Macs that used a 68HC05 program called “Egret” to run ADB and other functions. However, Egret’s 2.0 rewrite, “Cuda”, wasn’t working properly. I figured out what was wrong with it, and that enabled keyboard and mouse input for the Mac LC 520 and Color Classic machines, and it also opens the door to emulating the remaining 68030 machines and starting on the 68040s. (Cuda itself was used in the later 68030 Macs, most 68040 Macs, and all PowerPC Macs through the original iMac).
Many of the emulated NuBus video cards for the Mac were improved this year by Vas Crabb. A variety of cards got cleaner, better emulation and some cards with “virtual desktop” and pan/zoom features including the SuperMac Spectrum series got those features implemented. The Apple cards in particular got full CRTC emulation, allowing them to set the exact correct video modes. Also, the acceleration feature in the SuperMac Spectrum PDQ was finally fully supported, giving faster desktop updates.