Date : Fri, 19 Dec 2008 12:55:24 -0000
From : mfirth@... (Michael Firth)
Subject: IDE ADFS versions
----- Original Message -----
From: "Phill Harvey-Smith" <afra@...>
To: <bbc-micro@...>
Sent: Thursday, December 18, 2008 6:50 PM
Subject: [BBC-Micro] IDE ADFS versions
> Hi,
>
> I have just started using the RetroClinic IDE-CF adapter, tjis comes
> with two versions of ADFS that run on the BBC-B.
>
> v1.32 which will run out of sideways ram, and keep it's buffers there
> also, so leaving page unchanged. This version has no online help.
>
> v1.33 will also run out of swr, but also changes page, but does have
> online help.
>
> Does anyone know if the 1.32 help had to be removed to compress it
> enough to put it's buffers also in SWR ?
>
> Cheers.
>
> Phill.
>
> --
Having spent a while investigating this when I got my CF adapter, I can
confirm that the 1.32 image is a patched version of the ADFS 1.31E00 image
available at ftp://ftp.nvg.org/pub/bbc/rom/Acorn/fs/ADFS-1.31E00.rom. Though
this ROM is in the Acorn directory on that site, I'm not convinced its an
Acorn creation - certainly there are bits of the code that are significantly
different from the 1.30 / 1.50 standard Acorn ADFSes, while those two ROMs
are quite similar, apart from the BBC / Master changes. My best guess is
that this 1.31 ADFS is either an ACP or Solidisk creation, but I could be
wrong, and it doesn't have any company attribution (probably to save space).
Given that almost 4k (1/4) of the 1.31/1.32 code has been pruned out to
allow the workspace to be at the top of the SWR, I'm not surprised that the
help has gone. What was a bit of a surprise is that the *COMPACT command is
still present, but that *COPY isn't. If I were creating such a cut-down
ADFS, I would have ditched *COMPACT in favour of a disk based version, and
retained *COPY, which is probably less machine code than *COMPACT.
Despite what Mark said, I never did get the 1.32 version working on a Tube
enabled BBC, partly due to lack of time, partly due to not having a 2nd
processor to test with, and partly due to not having a deep enough
familiarity with how the ADFS code works to narrow down where the problems
are. Given that on a dual processor system the value of PAGE on the IO
processor isn't generally relevant, as most code runs on the 2nd processor,
and the 1.33 ADFS runs on a Tube system, and has more features than the 1.32
one, the best option is probably just to use the 1.33 version.
If anyone else has the time / energy to try and get the Tube support
working, the current BASIC patch files I have are at
http://www.firths.org/bbc. Alternatively, if anyone can point me at a PC
emulator that emulates both Tube and the IDE hardware, and has some form of
debugger built in, that might help me crack the problem at some point.
BeebEm has 2 out of the 3, but its ROM debugging could do with allowing you
to specify a ROM number to match - otherwise setting a breakpoint is a bit
tricky, as it will break any time any of the 16 ROMs hits that address.
For copying DFS to ADFS, I found the JGH patched version of 'TreeCopy' to
be the most useful, as it gives the option of copying the DFS directories as
files name 'X/YYYYYY', rather than creating single character directories,
which is easier to make work for some programs. If you don't have it, its
available in the HADFS support images on his http://mdfs.net/Software/HADFS
site.
Hope this is helpful
Michael