I'm ready to test F&I but I'm still having disk swapping issues.
Whether I swap (by pressing CTRL-SHIFT-F2) on the desktop or when requested in the game I just get an hour glass that never goes away.
The JFDs are just extracted from the zip with their original filenames to a folder on the SD card
Update on next release
Re: Update on next release
Have you removed the PC file extension? The files need to end with F10xxx01 and F10xxx02 etcringdings wrote:The JFDs are just extracted from the zip with their original filenames to a folder on the SD card
It's all explained in the ADFFS !Help file, but to save you looking, when swapping discs it looks for a file of the same filename as the one already mounted, but with the last character changed to the requested disc no. Hence why all the floppy images end with their unique ID.
Re: Update on next release
That's done the trick, thanks.
I did read the !Help quite a few times and took in that they needed to end 01, 02, etc, but didn't take in that you needed to remove the extension. That's the trouble with living in PC world for the last 18 years!
I did read the !Help quite a few times and took in that they needed to end 01, 02, etc, but didn't take in that you needed to remove the extension. That's the trouble with living in PC world for the last 18 years!
Re: Update on next release
You make a good point, I'll reword it so it specifically mentions that the extensions need removingringdings wrote:That's done the trick, thanks.
I did read the !Help quite a few times and took in that they needed to end 01, 02, etc, but didn't take in that you needed to remove the extension. That's the trouble with living in PC world for the last 18 years!
![Cool 8-)](./images/smilies/icon_cool.gif)
I did ponder for ages on what the filenames should be and decided that as the ZIP's were more than likely to be downloaded on a PC and probably used under emulation or transferred across, retaining the extensions would make it obvious as to which files were which. FTPc in combination with MIME extensions also does a good job of setting the filetypes automatically - provided MIME is configured of course. I should probably add a section on what needs adding to MIME.
There's a bug in ADFFS, as it should fail silently when it can't swap discs and not hang or leave the hourglass on as it does currently - it's already on my list of bugs to resolve by the looks of it (looking at the bottom of ChangeLog)
![Embarrassed :oops:](./images/smilies/icon_redface.gif)
Re: Update on next release
Finally managed to get to into my computer room to do some testing on physical StrongARM yesterday (the room's been full of kitchen and my son's room, whilst we're rebuilding both)
I didn't test many games, however what I did note was:
EDIT: Seems I'm already NOPing the Sound_Enable SWI in No Excuses, so it may not be DMA related.
I didn't test many games, however what I did note was:
- No sound in No Excuses and Rotor
- Lemmings crashed when starting a level
- Chuck Rock crashed at the prompt for disc 2
EDIT: Seems I'm already NOPing the Sound_Enable SWI in No Excuses, so it may not be DMA related.
Re: Update on next release
Fire & Ice has frozen 3 times on level 1 at different places.
Re: Update on next release
Same as I was seeing then. I think it's related to the cache flush issue I noticed on StrongARM on Friday.ringdings wrote:Fire & Ice has frozen 3 times on level 1 at different places.
Whilst recording the YouTube videos I also noticed Mad Professor Mariarti hangs randomly, seems to be a longer term issue though as I went back a few versions and was still seeing the problem. It would either hang or just jump back to the main menu for no apparent reason.
Listening to Pac-mania ...
... it sounds like one of the channels is not playing the sound. There is a bug in the channel handlers (actually it's a bug in RISCOS that was introduced with RO2 where the GateOn entry isn't called when a new sound is played) where the SCCB isn't initialized when a new sound is played, as a consequence it reads from page zero. I fixed this in the boot script by making it exit early if the SCCB value in question is zero - I wonder if it's that, that's causing the sound dropouts?
As this bug impacts most channel handlers (I'm currently fixing them in the boot scripts), I'm considering changing the voice handler entry code, so it tracks if the GateOn entry has been called prior to the Fill/Update entries being called.
EDIT:
Couldn't sleep, so quickly coded this up. It fixes Conqueror (possibly others, but haven't tested yet), but only partially resolves the problem with Pac-mania. It would appear the voice handler is being passed invalid sample addresses - which may explain why the samples don't play.I'm considering changing the voice handler entry code, so it tracks if the GateOn entry has been called prior to the Fill/Update entries being called.
Re: Update on next release
Full adffs249h ZIP on the dev site, which gets Poizone working. I've also put both the JFD image and complete archive of Poizone on the dev site.
We'll need to retest all currently working games as I've added support to hypervise SWI's called from BASIC. This was added whilst getting The Legend of the Lost Temple to work under the JIT, it's now running on RO3.71 SA emulation but not quite there on the Pi yet.
We'll need to retest all currently working games as I've added support to hypervise SWI's called from BASIC. This was added whilst getting The Legend of the Lost Temple to work under the JIT, it's now running on RO3.71 SA emulation but not quite there on the Pi yet.
Re: Update on next release
Updated obey.zip and ADFFS modules on the dev site, which get The Legend of the Lost Temple working.
Turned out to be either a bug in the code or a change in behaviour between RO3.71 and RO5 within OS_SpriteOp 36. I've implemented it as a change in behaviour as the RO3 PRM doesn't state what the correct parameters should be...the RO5 Wiki does however and it doesn't match what LotLT is doing.
Turned out to be either a bug in the code or a change in behaviour between RO3.71 and RO5 within OS_SpriteOp 36. I've implemented it as a change in behaviour as the RO3 PRM doesn't state what the correct parameters should be...the RO5 Wiki does however and it doesn't match what LotLT is doing.
Re: Update on next release
Testing with adffs249h and updated obey.zip:
-Lemmings: error on data transfer as soon as it's launched. Other time I got "ADFFS Exception: ADFFS5010 Branch through zero"
-Pacmania: hangs on loading picture
-Cannon Fodder: hangs after "Cannon Fodder Loading" text
-Blowpipe: seems to work fine
-2067 BC: seems to work fine
-Chuck Rock: hangs on Krisalis logo
-GODS: seems to work fine
-James Pond: seems to work fine
-Magic Pockets: seems to work fine
-Nebulus: seems to work fine
-Axis: seems to work fine
-Fire & Ice: seems to work fine
-Mad Professor: seems to work fine
-SWIV: hangs on black screen as soon as launched
-Lotus II: hangs on black screen after Krisalis logo
-Xenon 2: seems to work fine
-Rockfall: seems to work fine
-Paradroid 2000: seems to work fine
-Populous: seems to work fine
-Thundermonk: seems to work fine
-Heimdall: hangs on disk swap when I insert disc 2 to disembarc on an island
-Revolver: seems to work fine
-Lemmings: error on data transfer as soon as it's launched. Other time I got "ADFFS Exception: ADFFS5010 Branch through zero"
-Pacmania: hangs on loading picture
-Cannon Fodder: hangs after "Cannon Fodder Loading" text
-Blowpipe: seems to work fine
-2067 BC: seems to work fine
-Chuck Rock: hangs on Krisalis logo
-GODS: seems to work fine
-James Pond: seems to work fine
-Magic Pockets: seems to work fine
-Nebulus: seems to work fine
-Axis: seems to work fine
-Fire & Ice: seems to work fine
-Mad Professor: seems to work fine
-SWIV: hangs on black screen as soon as launched
-Lotus II: hangs on black screen after Krisalis logo
-Xenon 2: seems to work fine
-Rockfall: seems to work fine
-Paradroid 2000: seems to work fine
-Populous: seems to work fine
-Thundermonk: seems to work fine
-Heimdall: hangs on disk swap when I insert disc 2 to disembarc on an island
-Revolver: seems to work fine