Baltak Rampage, new MSX2 demo by Traktor (Revision party 2016)

Ascended (19463)

03-04-2016, 23:19

Can someone try Baltak Rampage on blueMSX with the "MSX2 - Philips NMS-8250" machine and check whether the torus part is OK? It's using the same disk ROM as the openMSX 8250.

By bore

Master (166)

03-04-2016, 23:32

Well, unfortunately the two 8280's we tried demo on belongs to Esau and Altruist.
I could ask Esau next time we have a computer meetup, but it might be a while. "Real life" stuff got put on hold while we tried to get Baltak finished. (Some things like the zoom in the beginning never got sorted out and remains a buggy mess, other things just got tweaked until it kinda worked.)
I could dump the ROM from my 8255 if you want it. The bug doesn't occur there either.

By bore

Master (166)

04-04-2016, 00:11

If you haven't found the spots yet the flashing of the palette in the torus part is done at $651F and $6548. The latter one being where the palette is restored.
The palette written is stored at $678A (Also used for the first part where the palette is correct.)
It looks like neither of the register 16 writes are protected by a di/ei pair so it could just be a matter of unfortunate timing but that wouldn't explain why you get different results with different disk ROM.

By Manuel

Ascended (19463)

04-04-2016, 07:14

Bore, that 8255 one is interesting too then

By Manuel

Ascended (19463)

04-04-2016, 21:26

OK, well, so it works with that diskROM. But it also works with the original diskROM if you change screen settings to KEY OFF: WIDTH 80.... (in the RTC).

So, I guess this would also be happening in a real MSX with that specific combination, probably indeed due to that missing di/ei. Bad luck with interrupts? Anything that influences that timing could then break or fix it.