Acorn Archimedes Software Preservation

[breach]

Would the new disc image format be compressed like APD? If it is, then maybe storing the blanks wouldn't really be much of a problem. How much difference is it likely to make? Some amount of bloat should be acceptable, I think. Maybe if it makes the disc image over 25% larger than it needs to be then it might be a problem.

[MarkYoungIW]

@Jon
Saw on Iconbar you were looking for the patches from:-

http://web.archive.org/web/200710090...nadoo.nl/tvdb/

I have spoken to Theo and the website is now at:-

http://members.upc.nl/t.boogaert6/

Hope this helps.

[MarkYoungIW]

The ADFFS record sounds impressive. Certain elements remind me of using Vine Micros Replay on the Beeb.

[Jon Abbott]

Originally Posted by breach

Would the new disc image format be compressed like APD? If it is, then maybe storing the blanks wouldn't really be much of a problem. How much difference is it likely to make? Some amount of bloat should be acceptable, I think. Maybe if it makes the disc image over 25% larger than it needs to be then it might be a problem.

Yes, the image files will be GZ files. The reason for keeping the image size down was to reduce the memory footprint of ADFFS. As I've done some major rewrites recently it's probably now in a state where I can leave the data compressed in memory and decompress on the fly. I can't do that with APD's easily, but could do it with the new format with little effort.

Originally Posted by MarkYoungIW

I have spoken to Theo and the website is now at:-

http://members.upc.nl/t.boogaert6/

Hope this helps.

Certainly does, thanks, I'll take a look.

Originally Posted by MarkYoungIW

The ADFFS record sounds impressive. Certain elements remind me of using Vine Micros Replay on the Beeb.

Spooky, I have one in front of me! I did consider dumping memory to disk, Replay style whilst playing Fire & Ice to the 3rd level repeatedly - god that was boring! Must have done it every 20 mins for the best part of a day, when Tom and I were diagnosing what was wrong with the APD!

[MarkYoungIW]

Originally Posted by Jon Abbott

Spooky, I have one in front of me! I did consider dumping memory to disk, Replay style whilst playing Fire & Ice to the 3rd level repeatedly - god that was boring! Must have done it every 20 mins for the best part of a day, when Tom and I were diagnosing what was wrong with the APD!

I've just found an apparently hacked copy of Fire+Ice in my some discs I have from the 90s, and there is a text file on one of the disks to say although hacked, level 3 has no ememies!!!

[Jon Abbott]

1.25 available which fixes the crashing when swapping discs. Download ADFFS

Now that I'm happy ADFFS is stable (pending bug reports), I'm going to make a start on the ripper. You'll see in the latest code that there's two commands related to recording, I need to expand them to support scatter list reads and then code the front end to perform it.

The back end will take less than a day...the front end meanwhile...lord only knows, I haven't written a BASIC WIMP app for 22 years!

[Jon Abbott]

My ideas on the floppy image file structure. If we can get this firmed up, I'll code it into ADFFS and put out some sample images. Hopefully, we can also get Fire & Ice re-imaged and working sooner rather than later.

Official files should be named "[optional text name] Fxxxxxnn" (see below for xxxx / nn explanation) eg "Fire & Ice (1995) (Warner Interactive Entertainment) F0000101" or shortened to "F0000101" on an Arc that doesn't support long file names. Provided the last char is the disc number, this could actually be anything, so "FireIce1" would be valid.

+0 ASCII "JFDI" - 4 chars for quick file type check during loading.
+4 Minimum version of ADFFS required to read eg 1.25 would be 125
+8 Memory allocation required to load file in bytes. ie uncompressed file size
+12 Disc sequence number - SSSSnnnn, where SSSS is the number of discs in this set, nnnn is the number of this disc.
+16 Game ID number (xxxxx) - official release ID of this floppy set. Unofficial releases should have an ID of zero.
+20 Image version number: VVVwwwww, VVV - major release version, 0 being initial release. wwwww - starts at 0 for each major release and is increased on each flush / save as to track local changes
+24 Offset to Track Table, the end of which is denoted by the start of the Sector Table
+28 Offset to Sector Table, end of which is denoted by the start of the Data Table
+32 Offset to Data Table
+36 Offset to optional Delta Track Table (used to store writes), 0 denotes not present
+40 Offset to optional Delta Sector Table (used to store writes), 0 denotes not present
+44 Offset to optional Delta Data Table (used to store writes), 0 denotes not present
+48 Disc title. eg "Fire & Ice (1995) (Warner Interactive Entertainment) disc 1 of 2", zero terminated.
+304 ... expansion space (currently 0 bytes)

[Track Table]
+0 Track 0 offset within Sector Table
+4 Track 1 offset within Sector Table
...
+n*4 Track n offset within Sector Table

[Sector Table]
+0 tt0dnnCs - Sector Header description (see below)
+4 Offset to start of sector data in Data Table
... repeats for all sectors in track, two doubles per sector
+n FFFFFFFF - end of track marker

[Data Table]
... X bytes of sector data in native byte format - identical to an ADF file.

Delta tables are identical to tables above and are checked first on each read/verify disc request.

Sector Header description:
tt - time of sector from start of index marker in milliseconds, max being 200
0 - currently unused nibble
d - Density value, ie 1 - single, 2 - double, 4 - quad, 8 - octal etc
nn - sector number
C - CRC bits, bit 0 - ID CRC invalid, bit 1 - Data CRC invalid, bits 2/3 unused
s - 1772 sector size, ie 0 - 128 bytes, 1 - 256 bytes, 2 - 512 bytes, 3 - 1024 bytes

[breach]

Originally Posted by Jon Abbott

1.25 available which fixes the crashing when swapping discs. http://adffs.filecore.net

Thanks for the update!

I just tried the High Risc Racing ADFs again. It doesn't crash any more when I change disc, but nothing else happens either. I just remain on the "Please Insert HRR 2" screen.

Edit: Exactly the same with Pesky Muskrats.

[Jon Abbott]

Originally Posted by breach

I just tried the High Risc Racing ADFs again. It doesn't crash any more when I change disc, but nothing else happens either. I just remain on the "Please Insert HRR 2" screen.

Edit: Exactly the same with Pesky Muskrats.

Quite a few wait for a mouse click or key press, but don't prompt as such.

[breach]

Originally Posted by Jon Abbott

Quite a few wait for a mouse click or key press, but don't prompt as such.

I already tried clicking and pressing return/space, in both games.

[Jon Abbott]

Originally Posted by breach

I already tried clicking and pressing return/space, in both games.

I'll see if I have them and try to find out what's going on.

EDIT: Pesky Muskrats - Waits in a loop with IRQ's disabled - hence, why you can't change discs via keys. Will have to put my thinking hat on to figure a way around that one!

High Risc Racing just locks the machine solid when I click on what I think is the race icon. Reading the ReadMe - "HRR may lock the computer requiring a hard reset" - nice. So...couldn't test it.

EDIT2: HRR - got it to at least prompt for disc 2 on Arculator. Same issue as Pesky Muskrats though.

[breach]

Ah, OK, at least there's only one problem, not two.

[Jon Abbott]

Originally Posted by breach

Ah, OK, at least there's only one problem, not two.

I believe I can work around it. I've just analysed Pesky Muskrats and its performing a MiscOp 1 (poll change), over 100 times a second. I'm going to guess that High Risc Racing will be identical, as any game requesting a disc will have to poll the drive for a change.

EDIT: Fixed in the next release, although it will no doubt be broken for something else now!

[breach]

Heh, I don't have many multii-disc images to test with, unfortunately.

[MarkYoungIW]

I can't get Cataclysm to load. Tried it both on the real hardware and Arculator and get to insert disc 2 where I change the disk, it briefly says 'Filecore does not understand this disc', then resets.

It doesn't run at all with Omniboot as it says it cannot move memory.

[breach]

Originally Posted by MarkYoungIW

I can't get Cataclysm to load. Tried it both on the real hardware and Arculator and get to insert disc 2 where I change the disk, it briefly says 'Filecore does not understand this disc', then resets.

Are you using real discs or images?

[Jon Abbott]

Originally Posted by MarkYoungIW

I can't get Cataclysm to load. Tried it both on the real hardware and Arculator and get to insert disc 2 where I change the disk, it briefly says 'Filecore does not understand this disc', then resets.

If you load disc 2 straight into ADFFS, do you also get the error - in which case, is the image corrupt?

[MarkYoungIW]

Both images 1+2 load OK into ADFFS. I have written the images to actual discs and these work fine. (Apart from not loading with Omniboot even as actual media).

Does seem something occurs when it gets to disc changing in ADFFS.

[Jon Abbott]

1.26 available, which should fix Pesky Muskrats and High Risc Racing. It will probably crash on physical kit though, due to FileCore re-entrancy. Under emulation it should be fine.

[Jon Abbott]

Originally Posted by MarkYoungIW

Both images 1+2 load OK into ADFFS. I have written the images to actual discs and these work fine.

Cataclysm could be corrupting the RMA - once disk one has loaded, ADFFS can't see the data on any discs subsequently mounted.

Everything works outside of Cataclysm.

EDIT: Looks like Cataclysm in installing a descrambler on OS_GBPB (or one of the associated vectors). If you copy it to your HD and run it, it will crash. If you then look at your HD contents, its scrambled until a reboot.

It's the N_W_Water module that does this.

EDIT 2: N_W_Water hijacks the following vectors:
FileV - OS_File
ArgsV - OS_Args
BPutV - OS_BPut
GBPBV - OS_GBPB
FindV - OS_Find

And descrambles anything going through them, without any checks on the drive its coming from - very poorly coded. If anyone has the original floppy, check if the module is on it - the wording in it reads like a hacker!

EDIT 3: Both the PROTECTOR and N_W_Water modules have been modified from their original form, we'll need to reimage the original into APD or the new format.

I have some ideas on how to get around protection that hijacks the IO vectors, which I'll add to the list of things to sort. Can we first confirm the N_W_Water module is on the original disc.

EDIT 4: PROTECTOR is a variation on the Gremlin protection module used in Pandora's Box.