home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
The Datafile PD-CD 3
/
PDCD_3.iso
/
utilities
/
utilst
/
zidefs090
/
ZIDEFS090
/
Updates
< prev
next >
Wrap
Text File
|
1994-10-03
|
6KB
|
133 lines
ZIDEFS/ZIDEFiler updates
0.81
- Filer : Bugfix, called Wimp_Initialise without 'messages-required' pointer.
- FS : Default of 1 drive now assumed if IDEFSDiscs set to 0.
- FS : *Configure IDEFSDirCache now accepts and displays trailing 'k'/'K'.
- FS : All *Configures now issue error if junk follows parameter value.
0.82
- FS : Bugfix, DiscOp 0 (verify) would corrupt r3. Wasn't very serious though.
- FS : Oh no ! Can't even rely on Acorn ! The PC emulator uses addresses above
8Mb for some reason, which makes OS_ValidateAddress reject them ! This in
turn failed DiscOp's 1 and 2, which check with OS_ValidateAddress. Took me a
whole afternoon to track down what the hell was going on ...
Therefore, I made the validation optional, which is a new 'soft' option per
drive (commands *IDEValidate and *IDENoValidate). Default for validation is
currently OFF, i.e. use *IDEValidate in your bootup sequence if you don't
intend to use the PC emulator and want the extra safety.
- Filer : Added 'Validate' option to the filer menu to toggle validation.
- FS : Changed name of SWI _DriveAccess to _DriveOptions.
0.83
- Filer : Bugfix, an erroneous instruction caused overwriting of module space
instead private workspace. Affected a mostly harmless area, except for
overwriting some of the finalisation code, which in turn caused 'Incarnation
not found' errors when re-loading the module from a file (i.e. luckily
unnoticable in everday use).
0.84
- FS & Filer : Source and internals restructured. No functional changes.
0.85
- FS : Improved drive compatibility, drive parameters are now read with an
Identify Drive command to the drive (instead of reading the boot record
from the drive). The info is used with Initialise Drive (which was never
right in the first place, it was Diagnose Drive because of a typing mistake,
eek !). As no actual disc access is needed to establish the drive's physical
parameters, this should improve drive compatibility (except for drives that
can't do Identify Drive (these shouldn't exist anymore these days) ...).
My new drive (a Western Digital WDA2250 (240Mb)), which wouldn't work with
0.84 (which is why I looked into the problem in the first place), now works,
and my older drive still works. NOTE : The old SF & PS software (1.14)
doesn't work correctly either with the WD drive !!!
0.86
- First version for UCS Enger (for a totally different 8-bit interface).
0.87
- Filer : Dismount and Verify didn't specify IDEFS. Oops.
- Filer : Free used ShowFree -fs IDEFS, but the ShowFree code is not in the
filer module (I used a seperate module). Now it is (as it should have been
in the first place !), so it can run the fancy 'Free' stand-alone.
- Filer : Disc icon is now on left of ADFS floppy drive icons (on request of
UCS Enger).
- Filer : Added NameDisc to iconbar menu (on request of UCS Enger).
0.88
- FS & Filer : Tidied up sources. Some code changes.
- Filer : Moved disc icon back to 'Other filing systems' priority. That's
where it actually should be.
0.89
- FS & Filer : now use assembler library, some code changed and/or tidied up.
0.90
- Almost entire rewrite of FS & Filer, several bugfixes and improvements.
This version needs a re-*Configure. Most notably the FS can now handle
partitions and has extra supporting software for this purpose. See Guide.
The most important (and some less important) fixes/changes/additions are
listed below. Note that error and SWI numbers have changed from 0.89.
- FS bugfixes :
- Fixed/improved error handling in some places (notably *Configure and
(SWI/low-level) DiscOp/MiscOp were a bit flawed and/or incorrect in
issuing errors).
- *Configure IDEFSDrive setting was never actually used (always drive 4
by mistake). Fixed.
- DiscOp option bit 6 (ignore Escape during operation) was erroneously
checked as bit 5. Fixed.
- Removed explicit selection of IDEFS on initialisation (*Configure
FileSystem setting should actually cause it to be selected by default).
- FS improvements/changes :
- Removed, added and renumbered errors. See Reference.
- Removed, added and renumbered SWI's to match more closely with ADFS/RamFS
offsets, which may benefit some software relying on FS's to have the same
offsets for similar SWI's. See Reference.
- Removed SWI _IdentDrive and _DiagnoseDrive. Their effects can be achieved
with a new SWI _Execute (although quite differently), which is a general
purpose IDE command list executor (see Reference).
- Added SWI _DriveInfo, only for internal use.
- SWI _DriveOptions used to return r1=-1 if an invalid drive was specified.
Now it returns an error.
- Most IDE operations now have a timeout, which ensures an error is returned
in case of bad IDE responses, instead of hanging the system.
- Implemented harddisc partitioning (see Guide).
- Implemented 'intelligent' initialisation, which should survive bad
configurations, bad harddiscs and bad partition tables.
- Improved drive compatibility. Most notably with a brand new WD 1210
which was used to test partitioning (this is a truly weird drive). Also
NEC D3756 should work ok now.
- 'Validation on' used to check with OS_ValidateAddress, which would
invalidate reading from ROM. The latter is now allowed (to enable, e.g.,
*Save's from ROM).
- Filer bugfixes :
- Now ignores clicks on 'Disc name' in top menu level (used to do *NameDisc
with '' as disc name, which errorred).
- All clicks now check ADJUST -> re-open (unknown selections used not to).
- Filer improvements/changes :
- Moved 'Read only' and 'Validate' menu items to 'Options' submenu.
- Made partitioning software. See Guide.
- Made a patch program for !HForm (on RISCOS 3 App2 disc), to be able to
initialise IDEFS partitions. None of the old SF & PS software needs to be
used anymore now. See Guide.
- ROM code bugfixes :
- Oops, used to corrupt r4 (due to bad programming example in RISCOS 2
PRM's).
- Check for 'bad address' was wrong, it did not account for the start of
the code space.
- Length of first chunk (description string) was wrong.
- ROM code improvements :
- Shrunk error messages a bit
- Squeezed loader code to minimum (bit faster too)