ADFFS 2.80 public beta

Discuss ADFFS development and download test releases
Post Reply
JonAbbott
Posts: 2938
Joined: Thu Apr 11, 2013 12:13 pm
Location: Essex
Contact:

Re: ADFFS 2.80 public beta

Post by JonAbbott »

Vanfanel wrote: Sun Jan 02, 2022 2:29 pm -Last Ninja: Fails to launch sometimes ("Attempt to use badly nested error handler (or corrupt R13)", and when it works, it hangs at quit via CTRL+SHIFT+F12.
I've spotted an issue that might be causing the hang when quit. Could you uninstall The Last Ninja, update the PackMan lists and then download it again and retry.

Could you also try quitting from the main menu by just pressing ESC.

EDIT: I've also spotted an issue with its start-up code, which might be causing the random launch failures, so please test that as well.
Vanfanel
Posts: 576
Joined: Mon Sep 16, 2013 12:01 am

Re: ADFFS 2.80 public beta

Post by Vanfanel »

@jon: After doing that, Last Ninja does not work at all.
ADFFS: Exception: Data Abort, etc...
JonAbbott
Posts: 2938
Joined: Thu Apr 11, 2013 12:13 pm
Location: Essex
Contact:

Re: ADFFS 2.80 public beta

Post by JonAbbott »

Vanfanel wrote: Thu Jan 27, 2022 12:14 am @jon: After doing that, Last Ninja does not work at all.
ADFFS: Exception: Data Abort, etc...
Interesting. I'm not sure what would be causing that. Could you try F12, then *GOS and manually run the game from the command-line so you can capture the Abort details. Use *SHOWREGS to get the register values - a screengrab will suffice
Vanfanel
Posts: 576
Joined: Mon Sep 16, 2013 12:01 am

Re: ADFFS 2.80 public beta

Post by Vanfanel »

As you will see from these, I get different errors on different tries.
They happen if I boot in SUPERVISOR mode, too. No difference, apparently.
Just different tries, same ADFFS, same !Ninja version... and different results.
Crazy thing to debug if you ask me :)
Attachments
photo_2022-01-27_19-53-10.jpg
photo_2022-01-27_19-53-10.jpg (121.15 KiB) Viewed 1800 times
photo_2022-01-27_19-52-57.jpg
photo_2022-01-27_19-52-57.jpg (99.3 KiB) Viewed 1800 times
photo_2022-01-27_19-52-45.jpg
photo_2022-01-27_19-52-45.jpg (104.61 KiB) Viewed 1800 times
JonAbbott
Posts: 2938
Joined: Thu Apr 11, 2013 12:13 pm
Location: Essex
Contact:

Re: ADFFS 2.80 public beta

Post by JonAbbott »

Thanks for the photos. Both crashes are within 26bit Modules and going by the randomness you’re seeing, I’m suspecting stack corruption caused by an interrupt re-entrancy issue. I’ll take a closer look at all the IRQ entry points in the game and it’s Modules.
JonAbbott
Posts: 2938
Joined: Thu Apr 11, 2013 12:13 pm
Location: Essex
Contact:

Re: ADFFS 2.80 public beta

Post by JonAbbott »

Nothing obvious jumps out at me, could you try the following to see if either pin the issue down:

Load !Ninja.FrontEnd in an editor and add the following line to prevent its function key handler code:

Code: Select all

165 !event_keys=&E1A0F00E
And if that doesn't work, try a REM at the start of line 190 to prevent the music tracker:

Code: Select all

190 REM SYS "ArcTrac_Volume",100:*GoBaby
Vanfanel
Posts: 576
Joined: Mon Sep 16, 2013 12:01 am

Re: ADFFS 2.80 public beta

Post by Vanfanel »

@Jon Tried both changes, and they both cause the game to freeze on launch.
JonAbbott
Posts: 2938
Joined: Thu Apr 11, 2013 12:13 pm
Location: Essex
Contact:

Re: ADFFS 2.80 public beta

Post by JonAbbott »

At what point does it freeze? Do you get to the main menu? Did you see the game title fader? Or did it freeze before both of those on a blank screen?
Vanfanel
Posts: 576
Joined: Mon Sep 16, 2013 12:01 am

Re: ADFFS 2.80 public beta

Post by Vanfanel »

JonAbbott wrote: Sun Jan 30, 2022 8:24 am At what point does it freeze? Do you get to the main menu? Did you see the game title fader? Or did it freeze before both of those on a blank screen?
It hangs on the tittle screen, it does nothing but show the tittle and the ninja eyes picture, statically until I powercycle the machine.
JonAbbott
Posts: 2938
Joined: Thu Apr 11, 2013 12:13 pm
Location: Essex
Contact:

Re: ADFFS 2.80 public beta

Post by JonAbbott »

v2.80r now available in the OP (and packaged for testers.)

Previously only a few values in Page Zero were handled by ADFFS via dedicated variables that were shared across tasks. This build adds the full 32Kb Page Zero block, allowing code to execute within Page Zero and the IRQ vector is now handled by calling it directly instead of tracking the branch address written to it.

It's possible this change might affect games that touch Page Zero. I've tested Lotus Turbo Challenge 2, which takes over the IRQ hardware vector and Oh No! More Lemmings which touches VULAcopy, but there will be others.
Post Reply