Recent content by Arbee

I've been highly skeptical of AI (and generally remain so), but the recent model versions (like in the last 3-6 months) have been a significant improvement. The Apple II guys recently got Claude to convert the Orca/C compiler for the Apple IIgs from Pascal to ANSI C. (Yes, the official C...

Regarding sync and interrupts: the Mac OS wants a VBlank IRQ from each video card, which it uses to update the mouse pointer "flicker free" when it's on that card's framebuffer. If a card doesn't have an IRQ, the pointer will freeze and not move forever when it goes onto that monitor. (The...

Claude's getting pretty impressive. Orca/C for the Apple IIgs, the official on-device compiler, has been source-available for a while now, but it's written entirely in Pascal (because that compiler already existed when they started on C). Claude successfully converted it into 100% C that...

The neck connector is keyed, so the neckboard would also rotate 180 degrees. That might require extending the leads to it. To put the picture back to normal, you'd have to either rotate the yoke 180 degrees so it's back to it's original orientation (but 180 degrees from the now-rotated CRT) or...

The real Q800 can't do 24/32bpp and the Monitors control panel won't let you select it. In Basilisk it's understandable that you can select it anyway because there's no hardware emulation, its all injected custom drivers. QEMU's DAFB emulation is minimal (they don't emulate the CRT parameters...

Would it be possible to flip the CRT 180 degrees so the anode's on the other side, or is the analog board in the way then? You'd have to rotate the yoke to get the picture back the right orientation then but there's nothing special about any particular orientation as far as the CRT is concerned.

Also worth noting: all of the Quadras except the AVs and 630 have mostly-compatible video and use the same driver, and that makes corner case bugs more likely. The Q700/900 ROM programs the wrong VRAM base and vertical resolution if the 512x384 monitor is connected. The Q800 and later ROMs...

It actually was nothing complicated, just an oversight in MAME's 680x0 differentiation.

Yeah. I had a similar problem with the '030 and it turned out that Apple was checking the internal details of the 680x0 coprocessor interface (there are coprocessor interface registers in the CPU address space). The coprocessor interface doesn't exist on the '040 so it's not that. The...

The CRT usually isn't the problem on TVs/monitors of that age. Bad caps seems much more likely.

Yes, it's known. Throwing an F-line trap on FPU instructions for the '040 gives a sad Mac error and I haven't investigated why that is so MAME always emulates a full '040.

I'm pretty sure I've posted this before, but let's do it again.

Most PCI PowerMacs have a hardware cursor, and I believe MacOS *does* use it. @joevt or @dingusdev would know more though.

I'd generally expect the real VIA to be slower than the pseudo-VIA given it's more complex, so I'd lean towards the faster CPU in the IIci making the difference. And yeah, I understand why FIFO full bit went away. It's not useful in any case, especially if you just leave the ASC enabled after...

Thanks again Callan! Wrenches are good, it makes sure we understand what's actually happening. That actually does line up with what I'm seeing in MAME where I got IIci perfect when spamming the Sound control panel, but SE/30 and friends are two IRQs short. Like Doug I don't have an...