The Darkness of Raven Wood

[pau1ie]

I guess the command you used was *CONFIGURE FILE n where n is a number from 0 to 15 corresponding to the ROM slot that the relevant filesystem ROM was in..?

Exactly

[lurkio]

The STOP error sounded familiar so I tried to recreate it using Elkulator. If I convert the SSD file to a UEF and try to run it, I get the same error. The code looks like it's doing things with the cassette filing system workspace, so if the DataCentre uses that to load files then it may well have problems.

The game definitely won't run off tape as it uses OSWORD &7F to do direct disc-sector reads to load the location graphics during gameplay.

The game does seem to run normally on .SSD in Elkulator though:

• 

  Screenshot of Raven Wood running on .SSD in Elkulator

[davidb]

The game does seem to run normally on .SSD in Elkulator though:

Yes, I was only testing whether it would run from tape, via a UEF file, as a step towards seeing if it would run from ROM.

Like a number of other disk-based adventures, it does some things that make that a bit difficult.

[daveejhitchins]

Yes, it would be a very nice addition to the MGC list.

Dave H

[davidb]

Yes, it would be a very nice addition to the MGC list.

Ideally, yes, but it would need to be modified to run there. It would also need to occupy a number of ROM banks.

[radiorama]

I've been following the development in the other thread and am impressed by the graphics and the teamwork that led to the game almost as much as by the game itself.

It works great on B-em, any chance of getting it to work on my Master Compact with ADFS? I get a STOP at line 560 error, guess that's due to the game directly accessing the disc sectors.

Carlo.

[lurkio]

It works great on B-em, any chance of getting it to work on my Master Compact with ADFS? I get a STOP at line 560 error, guess that's due to the game directly accessing the disc sectors.

Yes, the game relies on direct DFS sector reads with OSWORD &7F.

Are you able to load the DFS ROM into Sideways RAM in your Master Compact --

• http://www.8bs.com/submit/acornappspdf/053.pdf

-- and then try to use the existing version of the game?

[radiorama]

Yes, the game relies on direct DFS sector reads with OSWORD &7F.

Are you able to load the DFS ROM into Sideways RAM in your Master Compact --

• http://www.8bs.com/submit/acornappspdf/053.pdf

-- and then try to use the DFS .SSD disc image of the game?

Yes, I am. Actually, I have DFS burned into an EPROM inside my Compact. But I'm always looking for a new challenge

I just started tackling the problem and already hit the first stumbling block. The machine won't even try to access the disc when OSBYTE &7F is called from line 580. That's using ADFS. Any idea?

I think I could coax the thing into working. ADFS will already have a full catalog of the current directory in RAM as soon as it reads the !BOOT file. The plan is simple - read the starting sector of the image file from RAM and initialize S% from there. But it's doomed to fail as long as OSBYTE &7F doesn't agree to reading sectors from the disc...

C.

[tricky]

I think you could use the earlier version without all the images packed in to one file (pre compression, but you could add that back), it needs more files and more storage, but that is probably OK on a compact, but having one disk for all would be nice.

[lurkio]

I just started tackling the problem and already hit the first stumbling block. The machine won't even try to access the disc when OSBYTE &7F is called from line 580. That's using ADFS. Any idea? I think I could coax the thing into working. ADFS will already have a full catalog of the current directory in RAM as soon as it reads the !BOOT file. The plan is simple - read the starting sector of the image file from RAM and initialize S% from there. But it's doomed to fail as long as OSBYTE &7F doesn't agree to reading sectors from the disc...

I don't think OSWORD &7F works with ADFS:

• http://www.sundby.com/retro/copied_web_ ... 005723.htm

I think you could use the earlier version without all the images packed in to one file (pre compression

That version won't have the latest bugfixes (to game logic, etc.).

[radiorama]

I think I got it working with the Master Compact and ADFS. The attached file contains one DFS and one ADFS disk image with identical contents.

This version correctly detects the BBC Master Compact as well as the Master 128. When it comes to loading images from the big file, it checks whether DFS or ADFS is present and acts accordingly. I successfully tested it on:

* B-em in BBC Micro mode;
* actual Master Compact with DFS;
* actual Master Compact with ADFS.

I wouldn't mind some credits, should this version be officially released (PM me for real name).

Technical details on modifications follow.

I modified four files:

LOAD (BASIC): added INKEY-256 check for Master Compact as well as Master 128 on line 160. Added code to determine starting sector of big image file with ADFS on lines 560-600.

LOAD2 (EXEC): added INKEY-256 check for Master Compact as well as Master 128.

RW (BASIC): added alternate code for loading images from ADFS disks on lines 790-839.

RW2 (BASIC): added alternate code for loading images from ADFS disks on lines 326-332. This is still untested because I haven't played the game that far yet.

Best,
radiorama.

[Arcadian]

Thanks for adding the additional compatibility - nice to have a version that also works on my Compact!

I've tried using the ADFS image in Model B+ mode under BeebEm but get the STOP at line 600 error - assuming that there's enough memory to run the game (PAGE is at &1F00 by default on ADFS-enabled Beebs) might you be able to tweak the detection code to allow the game to run?

Likewise, I get the same STOP message when trying the ADFS version on an Electron with Plus 3 - again; if it's just the machine detection in the loader that's preventing it from loading, it would be fab if you could fix it up!

But if there isn't enough memory with the standard Electron ADFS (PAGE is at &1D00 on the Plus 3) perhaps it could be made to work with the PRES &E00 ADFS - which would load into Sideways RAM? (If you needed to test this,
then the old Beta 9b version of ElectrEm supports the &E00 ADFS roms - you just need to make sure you configure them as RAM (not ROM) on the Roms page under Options (F11).

[daveejhitchins]

perhaps it could be made to work with the PRES &E00 ADFS - which would load into Sideways RAM?

The MGC runs PAGE at E00 - That may be a good test bench

Dave H

[lurkio]

... assuming that there's enough memory to run the game (PAGE is at &1F00 by default on ADFS-enabled Beebs) might you be able to tweak the detection code to allow the game to run?

PAGE might well be a problem, but it sounds like the game isn't getting far enough for us to know for sure! The game is erroring early on, at the stage where it tries to find the starting sector of the big file full of picture data.

Though I don't fully understand the details, the problem seems to have something to do with the way Carlo has implemented the starting-sector search in $.LOAD: I think that the implementation assumes that a copy of the disc catalogue will be in RAM somewhere above &C000, which I believe it will be on a Master and/or Compact, but probably not on a B Plus..? (I might be quite wide of the mark here though!)

The MGC runs PAGE at E00 - That may be a good test bench

La la la la la I can't hear you...

[Arcadian]

Hi Chet, thanks for the reply - just to clarify I'm curious if the game will run on any Model B/B+ with ADFS, not just the B+ specifically (I used B+ mode in BeebEm as it offered a quick and convenient way of emulating a Model B series machine running ADFS).

Like you say it might not be possible due to PAGE etc BUT if it's merely the detection routines in the loader that are preventing it running on a B/B+ (or an Electron Plus 3/Electron Plus 3 with PRES &E00 ADFS) it would be good to get it remedied.

I guess it would also be worth seeing if the game runs/can be made to run on a Model B/Master with Winchester or CF-Flash hard drive ...

[radiorama]

@Arcadian: I'm working on it

C.

[radiorama]

Here is a new version adapted to different incarnations of ADFS. Tested and apparently working with:

* B-em in BBC B mode;
* B-em in BBC B+ mode with ADFS;
* B-em in Master 128 mode with DFS;
* B-em in Master 128 mode with ADFS;
* ElectrEm with standard ADFS;
* ElectrEm with PRES E00 ADFS;
* actual Master Compact with DFS;
* actual Master Compact with ADFS.

I've never been around an Electron, so I've had to learn the hard way that, on the Electron, ROMSEL is at &FE05, not at &FE30, and has to be driven in a very specific way.

The "LOAD" program had to be split in two, otherwise it wouldn't fit in the available memory with a 1F00 ADFS while in MODE 1.

If ADFS is detected, the program looks for the ADFS signature "Hugo" at &C401 (where the Compact ADFS stores the current directory information) and at &1201 (where regular 1D00 and 1F00 ADFSs store it); failing that, it systematically searches all the sideways RAMs/ROMs (that's where the PRES E00 ADFS stores the directory info). When a directory is found, a search starts for a "Lpics2" file and, if found, its starting sector on disc is finally read.

No further modifications were needed for "RW" and "RW2".

As far as I can say, ADFS pushing PAGE way up shouldn't be a problem. A 12 kB BASIC program running in MODE 5 with HIMEM at &5800, even with PAGE at &1F00, leaves abouk 2 kB of wiggle room.

IIRC, Winchester drives use ADFS, so I wouldn't be surprised if the game worked out-of-the-box with one of those.

Can someone confirm that "RW2" works correctly? I still haven't played the game that far (hacking it is at least as fun as playing it, and that's not to say that the game isn't great), therefore I'm still not sure that I modified the program properly.

-- Carlo.

[lurkio]

Here is a new version adapted to different incarnations of ADFS

Great work!

If ADFS is detected, the program looks for the ADFS signature "Hugo" at &C401 (where the Compact ADFS stores the current directory information) and at &1201 (where regular 1D00 and 1F00 ADFSs store it); failing that, it systematically searches all the sideways RAMs/ROMs (that's where the PRES E00 ADFS stores the directory info). When a directory is found, a search starts for a "Lpics2" file and, if found, its starting sector on disc is finally read.

I don't really understand ADFS, so I'm not sure if this would be possible, but couldn't you have just read the catalogue sectors of the disc into a free patch of RAM and then parsed that copy of the catalogue to find the Lpics2 starting sector? (That's what I did for DFS.)

[radiorama]

I don't really understand ADFS, so I'm not sure if this would be possible, but couldn't you have just read the catalogue sectors of the disc into a free patch of RAM and then parsed that copy of the catalogue to find the Lpics2 starting sector? (That's what I did for DFS.)

Unlike DFS, where the directory catalogue always resides on a specific sector of the disc, ADFS can have subdirectories. That is, the root is always in the same spot on disc, but subdirectories can be stored anywhere. And I like my games in subdirectories on compilation discs.

I'm not sure how I'd go about finding the starting sector of the current directory, so I followed this approach of using the RAM copy. Little did I realize there were so many versions of ADFS, each with a different hiding place for its directory image. Moreover, the code could be optimised by doing the file search in ASM or converting it to a BASIC FN (as it is now, there are three almost identical chunks of code that search for the image file inside the directory image).

BTW, a funny thing I noticed is that the Electron seems unable to display the colour yellow (or at least, ElectrEm is). On ElectrEm, colour 3 is white, colour 7 is gray. Is this the way the Electron display works, or is it an ElectrEm bug?

C.

[Arcadian]

Wow, thanks Carlo - really appreciate you taking the time to update the images.

I've quickly checked the ADL image under BeebEm in B+ mode with ADFS and also Elkulator in Plus 3 ADFS mode and in both cases the game indeed loads up fine (though I didn't test the game beyond a couple of locations hence can't vouch for RW2).

However I did try copying the files onto an emulated SCSI hard drive under BeebEm and though it would load the menu screen and display the instructions (option 1) if I tried loading the game (option 2) it would just hang on a flashing cursor (it did the same in both Model B+ mode and Master 128 mode). I've attached a hard disc image of a scsi drive containing the Ravenwood files for reference.

I'll try and get a Datacentre BBC B with CF/Flash ADFS hard drive set up later today to test it as well ... will let you know how I get on!

[radiorama]

However I did try copying the files onto an emulated SCSI hard drive under BeebEm and though it would load the menu screen and display the instructions (option 1) if I tried loading the game (option 2) it would just hang on a flashing cursor (it did the same in both Model B+ mode and Master 128 mode).

Strange... it works here under Beebem 4.14 in both Model B+ and Master 128 modes. I copied your hd image in my user folder, enabled SCSI, then CTRL+A, *dir ravenwood, CH."LOAD", it loaded up fine and I visited a couple locations.

C.

[Arcadian]

It's working ok here now as well!

The only thing I did differently originally was use *EXEC !BOOT as opposed to CHAIN "LOAD" but *EXECing the boot file is also now working for me as well ... hmm!

[Arcadian]

Ok, I did some further testing relating to CF-Flash IDE HDD operation on my Datacentre-equipped BBC B & BBC Master, below are the results (both were supplied fairly recently by Mark so should contain latest firmware).

BBC Micro Model B with Datacentre

ADFS 1.32 (PAGE at &E00) called by *XADFS
Runs from CF-Flash ADFS Hard Drive: Yes
Runs from Floppy Drive (Drive 4/5): Yes

ADFS 1.33 (PAGE at &1F00) called by *YADFS
Runs from CF-Flash ADFS Hard Drive: Yes
Runs from Floppy Drive (Drive 4/5): No ("Not found on channel 57 at line 510" after selecting option 2 to load game)
Edit: these results were typed up the wrong way around - see updated post below!


BBC Master with Datacentre

MOS 3.5 with ADFS 2.05
Runs from CF-Flash ADFS Hard Drive: Yes
Runs from Floppy Drive (Drive 4/5): Yes

MOS 3.2 with ADFS 1.53
Runs from CF-Flash ADFS Hard Drive: No (hangs on flashing cursor after selection option 2 to load game)
Runs from Floppy Drive (Drive 4/5): No (hangs on flashing cursor after selection option 2 to load game)

I might drop Mark H a line about the above issues (unless anybody else has any suggestions?).

Eventually I also intend to test the files on a BBC Micro and BBC Master without Datacentre, using various SCSI Winchester drives.

But next I'll move onto testing on actual Electron hardware (including Acorn Plus 3, Acorn Plus 3 with Prime DFS/&E00 ADFS upgrade, PRES AP3, PRES AP4, Slogger Pegasus 400, Slogger SEDFS, Cumana DFS (or maybe I won't bother with this one), Solidisk DFS/ADFS and Solidisk Winchester ... phew!).

Oh, and whilst on the subject of the Electron, once we've got an ADFS version of the game which we consider to be 'final' it probably makes sense to issue it as a non-interleaved 320k image (as opposed to 640k/ADFS "L" format) as the Acorn Plus 3 was only ever supplied with a single-sided drive (despite what the manual may say!) hence an ADFS "M" image could be written and used directly with a Plus 3 (as well as the Master Compact etc).

[jms2]

BTW, a funny thing I noticed is that the Electron seems unable to display the colour yellow (or at least, ElectrEm is). On ElectrEm, colour 3 is white, colour 7 is gray. Is this the way the Electron display works, or is it an ElectrEm bug?
C.

That sounds very strange! Electrons certainly can display yellow, and i wouldn't like to say it is an ElectrEm bug either because it's not one that I've ever noticed and it sounds a bit unlikely.
Thomas Harte the author does post on here, so maybe he can explain what may be going on.

[radiorama]

Ok, I did some further testing relating to CF-Flash IDE HDD operation on my Datacentre-equipped BBC B & BBC Master, below are the results (both were supplied fairly recently by Mark so should contain latest firmware).

ADFS 1.33 (PAGE at &1F00) called by *YADFS
Runs from Floppy Drive (Drive 4/5): No ("Not found on channel 57 at line 510" after selecting option 2 to load game)

This most certainly refers to line 510 of "RW", which OPENINs two files and reads some variables from them. This operation is definitely not among the most unconventional that are performed as the game loads, and shouldn't pose a problem.

BBC Master with Datacentre

MOS 3.2 with ADFS 1.53
Runs from CF-Flash ADFS Hard Drive: No (hangs on flashing cursor after selection option 2 to load game)
Runs from Floppy Drive (Drive 4/5): No (hangs on flashing cursor after selection option 2 to load game)

When the prompt to display instructions or load the game appears, the search for the image file on disc is successfully completed. Do you see a "thick", "MODE 6" kind of cursor? Does it ever go away?

In non-working cases, what does "P. INKEY-256" return?

C.

[Arcadian]

Ok, I did some further testing relating to CF-Flash IDE HDD operation on my Datacentre-equipped BBC B & BBC Master, below are the results (both were supplied fairly recently by Mark so should contain latest firmware).

ADFS 1.33 (PAGE at &1F00) called by *YADFS
Runs from Floppy Drive (Drive 4/5): No ("Not found on channel 57 at line 510" after selecting option 2 to load game)

This most certainly refers to line 510 of "RW", which OPENINs two files and reads some variables from them. This operation is definitely not among the most unconventional that are performed as the game loads, and shouldn't pose a problem.

I took a couple of snaps whilst testing the above:

Basically when I did a *CAT after the program bombed out, it catalogued Drive 0 (i.e. the first CF-Flash hard drive) instead of floppy drive 4.

And then when I pressed BREAK, ADFS was gone altogether!

BBC Master with Datacentre

MOS 3.2 with ADFS 1.53
Runs from CF-Flash ADFS Hard Drive: No (hangs on flashing cursor after selection option 2 to load game)
Runs from Floppy Drive (Drive 4/5): No (hangs on flashing cursor after selection option 2 to load game)

When the prompt to display instructions or load the game appears, the search for the image file on disc is successfully completed. Do you see a "thick", "MODE 6" kind of cursor? Does it ever go away?

Yes, that's the one. It just carries on flashing...

In non-working cases, what does "P. INKEY-256" return?

Haven't checked yet but will do so today and let you know!

Oh, and this morning I successfully tested the latest images on all of the Electron disc interfaces I currently have to hand, which are:
- Acorn Plus 3
- Acorn Plus 3 with Prime's upgrade, in both DFS and &E00 ADFS modes
- PRES AP3 (which had the PRES ADFS 1.11 rom installed)
- PRES AP4 (with PRES DFS 2.20 rom)
- Slogger Pegasus 400 (with both v1.01 and v1.02 of the Pegasus DFS roms)

As I mentioned above, I'll try it out with some of the more esoteric interfaces at some point, though it would appear that the game works on all of the most popular disc interfaces for the Elk, which is excellent news! Thanks again for your help here!

[Arcadian]

Ok, I've done a little more testing, it seems I may have got the results of the 1.32/1.33 ADFS (on the BBC B) the wrong way around on the previous post (oops) but I /have/ double-checked the results below.

In each case I've tried loading the game with both *EXEC !BOOT and CHAIN "LOAD".

BBC Micro Model B with Datacentre

ADFS 1.32 (PAGE at &E00) called by *XADFS
Runs from CF-Flash ADFS Hard Drive: Yes
Runs from ADFS Floppy Drive (Drive 4/5): No
"Not found on channel 57 at line 510" message after selecting option 2 to load game. PRINT INKEY-256 gives -1

ADFS 1.33 (PAGE at &1F00) called by *YADFS
Runs from CF-Flash ADFS Hard Drive: Yes
Runs from ADFS Floppy Drive (Drive 4/5): Yes

BBC Master with Datacentre

MOS 3.5 with ADFS 2.05
Runs from CF-Flash ADFS Hard Drive or ADFS Floppy Drive (Drive 4/5): Yes

MOS 3.2 with ADFS 1.53
Runs from CF-Flash ADFS Hard Drive or ADFS Floppy Drive (Drive 4/5): No
If *EXEC !BOOT used, it hangs on a flashing cursor after selection of option 2 to load game.
If CHAIN "LOAD" used, the game loads in but immediately displays "Channel on channel 57 at line 1210" after the What Now prompt. PRINT INKEY-256 gives 253 (see photo below).

[radiorama]

Thanks for checking, Arcadian. The problem might well have to do with an ADFS bug. According to this thread:

viewtopic.php?t=12500#p159797

IIUC, race conditions happen due to the way ADFS manages shared workspace in RAM. This would explain the mixed results with datacentre-enabled machines, where the outcome depends on how you load up the game, and the intermittent success you experienced with SCSI on BeebEm.

Unfortunately, in these cases there's no clear solution. One instruction properly placed might save the day (and, possibly, break something else). Does anyone have any experience dealing with this sort of problems?

C.

[Arcadian]

FYI I heard back from Mark, he wasn't able to shed any light on the quirks but suggested I ask resident ADFS guru, JGH ... who is now on the case! Fingers crossed Jonathan can get to the bottom of it!

[Dethmunk]

Hey guys..... OMG!!! You're still working on Raven Wood?!? This is amazing. I never realized you were all so passionate about it. This is awesome. It was just a hobby project to me and you guys have taken it and run... I can't believe you got it all working off physical media. Brilliant work to everyone.

Colour me gobsmacked!!