ADFFS 2.54
Re: ADFFS 2.54
@ArcEm is not a valid test case for this imho, since it does not sync to physical refresh rate and there's always tearing.
Re: ADFFS 2.54
I've just tested JP2/JP2+ (in std video mode) and it doesn't look like ArcEM implements VIDC correctly, so is no use in this case anyhowVanfanel wrote:@ArcEm is not a valid test case for this imho, since it does not sync to physical refresh rate and there's always tearing.
Testing on a physical A310 the game does not have smooth horizontal scrolling. Recording JP2+ running under the debug build at 240fps, the horizontal scrolling doesn't always scroll by 2 pixels, it's sometimes 4 pixels, hence the juddery scroll. I can only guess that the original game may have been designed for NTSC and microstutter has been introduced whilst matching the original gameplay speed at 50Hz.
Re: ADFFS 2.54
I've put what will hopefully be the final build full archive of 2.55 on the dev site. The loader sequence has now been modified to determine if the CPU supports rotated loads and runs ADFFS500 / ADFFS550 accordingly. It also checks that RO5 is a High Vector build and reports an error if not.
If the Pi is set to ARMv5 compatibility mode, the CPU will handle rotated loads and if it's set to ARMv7 strict mode, ADFFS will emulate rotated loads. If the CPU is in ARMv7 fast mode, the result will be unpredictable.
If the Pi is set to ARMv5 compatibility mode, the CPU will handle rotated loads and if it's set to ARMv7 strict mode, ADFFS will emulate rotated loads. If the CPU is in ARMv7 fast mode, the result will be unpredictable.
Re: ADFFS 2.54
@Jon: I was doing some tests yesterday and I will do more this weekend. It seems to be working great so far for the compatible games, but Ibix the Viking (I believe you wanted to include it in this release) is still running too fast on the Pi.
Re: ADFFS 2.54
The titles below currently run too fast as they don't regulate their speed. This isn't something I've found a workable solution for yet, but it doesn't stop them being release, as they work okay on true Archimedes:Ibix the Viking (I believe you wanted to include it in this release) is still running too fast on the Pi.
F10490 Champions: World Class Leader Board (1992) (Krisalis Software)
F10166 F.R.E.D. (1993) (Software 42)
F10202 HoverBod (1988) (Minerva)
F10204 Ibix the Viking (1989) (Minerva)
F10524 Letrouve (1993) (Les Images Bleues Foncées)
F10242 Manchester United (1990) (Krisalis Software)
F10487 Play It Again Sam 3: Letrouve (1993) (Superior Software)
F10285 Populous (1992) (Krisalis Software)
F10343 Sporting Triangles (1990) (CDS Software)
F10348 StarTrader (1989) (Gem Electronics)
F10265 Triumvirate I, Das (1991) (Uffenkamp Computer Systeme)
F10641 Triumvirate II, Das (1991) (Uffenkamp Computer Systeme)
F10393 World Class Leaderboard (1988) (Krisalis Software)
I've been doing more testing on physical Acorn's today and spotted a few fatal issues which I'm currently fixing. There's an issue with the sound initialization that causes a crash when ADFFS is first loaded (this also affects the Pi) and an issue with some games going into an invalid video mode on the RiscPC. They're minor bug fixes, so won't require any major retest.
Some games that should work on StrongARM aren't, which I mentioned previously. It's going to take weeks to track down the bug(s) so I'll release as-is, with the caveat that some games aren't working on SA. I've already spent two weeks trying to pin the issue down and not really got anywhere.
Re: ADFFS 2.54
@Vanfanel: I've put today's build up, the latest obey.zip and Fine Racer. Could you playtest it and see if its working. You have to collect all the numbered checkpoints before you can complete the level (even though the AI cars don't seem to collect them ) and get to the end before the other cars do. As with most Eterna titles, it's damn near impossible so I gave up testing it last night.
I'll add this to the release list if it looks okay.
Ironically this floppy causes ADFFS to hang on RO3.x, which is another bug I'm looking at currently in case it could also affect the Pi build.
EDIT: The "FIRE" key is @ or ' not / or ? as detailed in the instructions.
) and get to the end before the other cars do. As with most Eterna titles, it's damn near impossible so I gave up testing it last night.I'll add this to the release list if it looks okay.
Ironically this floppy causes ADFFS to hang on RO3.x, which is another bug I'm looking at currently in case it could also affect the Pi build.
EDIT: The "FIRE" key is @ or ' not / or ? as detailed in the instructions.
Re: ADFFS 2.54
Turned out to be a change I made yesterday that broke RO3.x support entirely!JonAbbott wrote:Ironically this floppy causes ADFFS to hang on RO3.x, which is another bug I'm looking at currently in case it could also affect the Pi build.
Whilst soak-testing Rockfall today, I noticed it crashes if left on the demo at the same place. I'll see if I can track it down, its possibly a Page Zero access or a bug.
Re: ADFFS 2.54
Further testing of games yesterday highlighted more bugs, which are fixed in today's build. I've also updated obey.zip with the Boot Script for Big Bang.
Re: ADFFS 2.54
One particular level in the demo triggers a Page Zero access in SoundDMA at line 3446. SoundDMA's pointer to it's RMA space (stored @ FAFF4034) is mysteriously zero. In fact most of SoundDMA's private space is zeroed out.JonAbbott wrote:Whilst soak-testing Rockfall today, I noticed it crashes if left on the demo at the same place. I'll see if I can track it down, its possibly a Page Zero access or a bug.
Re: ADFFS 2.54
@Jon: Fine Racer seems to work well on the Pi. It's hard but playable and it seems fine to me.
Big Bang works great too!
Any chances to see Hamsters?
Big Bang works great too!
Any chances to see Hamsters?