forums.jaspp.org.uk

@John: How comes Elite works again?? Didn't it need WIMP support?? Now it works!! Wow!
The only problem is that it hangs on CTRL+SHIFT+F12 quit. That being fixed, the game seems perfectly playable!

Vanfanel wrote:@John: How comes Elite works again?? Didn't it need WIMP support?? Now it works!! Wow!
The only problem is that it hangs on CTRL+SHIFT+F12 quit. That being fixed, the game seems perfectly playable!

Hmm, not sure how it's working, it wasn't when I accidentally fired it up a few weeks ago.

As its a WIMP app, CTRL-SHIFT-F12 will cause a lock up. You'd have to exit to the desktop (F12 I think), then exit the WIMP app to close it cleanly, that may also cause a lock/crash though as the sound Modules may crash when terminating.

Vanfanel wrote:@Jon: What was the parameter to set so ADFFS does not try to guess vsyncs/second and simply runs on whatever is being used? I have forgotten and we have written SO much in this forum that I can't find it...

It was stripped out long ago as ADFFS no longer counts VSync's. Frame swapping is done at physical VSync based on a PLL, so will always achieve 50Hz regardless of the hardware refresh rate. If the hardware is configured for 50hz, it will swap every VSync.

@Jon: That's not totally true. I found the enviroment variable in the docs:
set ADFFS$50Hz true

..and it DOES make a difference, even with 2.61 beta. Please don't remove it!. Whatever it's doing, it's doing very well: Pac-Mania has occasional hiccups without this (NOT frequent, but you know I can detect a lost/desync frame even if it happens every 2-3 minutes), and is PERFECT with it: no frame desync/lost to be seen.
It's no placebo effect, you can be sure of that

Vanfanel wrote:It's no placebo effect, you can be sure of that

Who'd have thought the placebo effect works on code

Updated Modules on the dev site, it was randomly corrupting memory when changing MODE and failed to set the VIDC20 vertical registers when a VIDC Type 3 MODE was requested. This resulted in no mouse cursor in Caverns, so no ship/score. A side-effect of fixing it, is the aspect ratio in Raw Power is now incorrect; I need to look at the resolution ratio and possibly tweak the code that works out when to line double.

In an attempt to resolve the random micro-stutter issue you're seeing in Pac-mania, I've shifted the PLL Vsync 50% out of phase with the physical VSync. I'm not certain this will resolve the issue, as you can never guarantee timer driven code will occur at the time you specify on RISCOS. I believe the issue is being caused by a delayed timer event, which is causing a frame to be dropped.

@Jon:

I have been testing the new modules, and I see no more micro-stutters in Pac-mania, even without ADFFS$50Hz.
However, if the ADFFS$50Hz functionality is still there, *please* don't remove it (I haven't understood if it's there or not, but there was a perfectly visible difference with it). It can be VERY useful for future comparisions.

Also, I got a wireless mouse which seems to be affected by ADFFS: it works perfectly well until I enter an ADFFS game and then quit via CTRL+SHIFT+F12. Then, the mouse cursor moves by itself to the lower-right corner of the screen.
Note the game (whatever ADFFS game I load) is perfectly playable. The problem occurs only after I quit the game.
Native games like Zool, Twinworld, Doom... don't have this effect, only ADFFS game. Any ADFFS game, actually. Any idea of what ADFFS could be doing at exit that could be causing this?

PD: Elite has stopped workin again

Vanfanel wrote:I got a wireless mouse which seems to be affected by ADFFS: it works perfectly well until I enter an ADFFS game and then quit via CTRL+SHIFT+F12. Then, the mouse cursor moves by itself to the lower-right corner of the screen.

Any idea of what ADFFS could be doing at exit that could be causing this?

Everything is shut down when a game quits, try RMKill'ing the ADFFS Module when it occurs.

Lemmings 2 is broken, I thought I'd fixed the VIDC Type 3 issues, clearly not.

Could you conduct two further tests after a reboot and let me know if either causes the mouse to go rogue.

1. F12, then enter:
   GOS
   GOARM3JIT 0
   MODE 13
   GOARM3JIT
   QUIT
   Press ENTER to return to the desktop and check the mouse
2. F12, then enter:
   GOARM3JIT 0
   GOARM3JIT
   Press ENTER to the desktop and check the mouse

Vanfanel wrote:if the ADFFS$50Hz functionality is still there, *please* don't remove it (I haven't understood if it's there or not, but there was a perfectly visible difference with it).

It was removed during the 2.60 testing, as the PLL based Vsync will trigger a blit every Vsync. I'll look at what's involved in putting it back, its not as straightforward as you'd think.

Vanfanel wrote:Elite has stopped working again

It was probably working due to luck, the only way it could possibly work would be if there were no other programs running at the desktop to trigger task switching. The localised APCS-R Module is what's breaking it, or more specifically the lack of a 26bit RMA when at the desktop.

@Jon: About the mouse tests, I have tried:
1- Killing the ADFFS modules when the mouse has already gone havok: The mouse is sill havoc with this.

2- This sequence of commands: This sequence fails because "MODE" is not found (error &d6). That's with ADDFS loaded.

3- This sequence of commands: This sequence does not cause the mouse to go havok: it keeps working as normal.

Sorry, MODE is an alias that's disabled by default. Use "ECHO <22><13>" instead.

On a separate issue, could you see if Magic Pockets still works.