home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Aminet 25
/
Aminet 25 (1998)(GTI - Schatztruhe)[!][Jun 1998].iso
/
Aminet
/
disk
/
misc
/
xfs.lha
/
xfs
/
xfs.readme
< prev
next >
Wrap
Text File
|
1998-04-26
|
7KB
|
162 lines
Short: V2.12b3 Multi filesys+.device, pc+mac+..
Uploader: msw@blackpool.ac.uk
Author: frans (francis swift)
Type: disk/misc
Replaces: xfs.lha
XFileSystem Version 2.12 beta release 3.
Consisting of xfsd (a disk handler for your L: directory)
and fd.device / file.device (for your devs: directory).
Current xfsd status:
readdir read write delete rename readpart format
ql + + + + +
spectrum + +
msdos + + + + + +
archimedes + +
amiga + + + + + +
minix + +
mac + + +
cpm + +
1541 + +
Current fd.device status:
read write format
Amiga tracks + +
Diskspare +
mfm (pc etc) + +
Current file.device status:
read write format
flat files + +
dms files +
pctask/janus hardfiles + +
As you can see, there is currently no way to format disks.
In fact, most of the code to format disks marked as writeable
exists, it's just there's no way to select it at the moment.
Actually, TD_FORMAT does work on fd.device, but only on disks
already formatted (it effectively just writes).
See the xfs.guide 'usage' section for further info.
-------------------------- xfs212b3 -----------------------------
Just a slight update to file.device. I re-wrote the xDMS source
to be re-entrant / thread safe ('pure' in amiga-speak), so now
file.device, which uses the decompression routines from xDMS,
can safely be opened multiple times for multiple DMS files.
I'll probably upload the modified xDMS source with the next
release of xfs.
-------------------------- xfs212b2 -----------------------------
There was a slight bug in the previous file.device when using
PCTask hard disk files. Basically, writing could trash them.
This was a bug introduced during re-organisation to allow proper
integration of DMS file access. Anyway, it should all work now,
reading/writing PCTask hard disk files (apparently the same format
as janus hard disk files), and reading DMS files. To implement
the DMS facility I used the source from xDMS (by Andre de la Rocha)
to replace my original routines (which called DMS externally).
The current xDMS decompression source is still very similar to
the original LhA decompression routines (written by Masaru Oki
and Haruyasu Yoshizaki), which are not re-entrant, so at present
you can only have one DMS file in use by file.device (ie only one
unit of file.device can be accessing a DMS file - other units can
still access other types of disk file). I have recently re-written
these LhA decompression routines (for another project) to make them
re-entrant, so the next release of file.device should allow any
number of units to access DMS files. Hopefully, it should soon be
possible to transparently access .ADZ files as well.
There is a mis-match between the current version of the devio
utility and the current release version of fd.device. This only
affects the -geometry option, and is caused by a mis-match in the
private structure used to pass back the low-level info (about
sector numbering and skew factors). This will be fixed next time.
Ther was a slight bug introduced by internal rationalisation of
the utf-8 / unicode routines which caused QL comments (which hold
task space info) to be messed up. Should be ok now.
-------------------------- xfs212b ------------------------------
Changed file.device to allow it to skip the first sector of PCTask
hard disk files, so you can use FILE0: (xfsd / file.device) on them.
Also new DOCS for file.device in the guide (->usage->file.device).
Oh yes it also sends back a sensible geometry for .d64 files. This
was necessary to implement a C64 disk handler.
1541 disk images may now be accessed via FILE0: (xfsd), but only to
display directories at present. You will of course require Catweasel
to access the disks directly. This is intended to eventually become
a combined 1541/1581 handler (when I get some 1581 info).
devd (the DEV: handler) can now accept a third parameter for FLAGS
after the unit number (ie DEV:multidisk.device,0,7) to enable
you to use it with Catweasel. Oh yes, devd is now mentioned in the
guide (->usage->devd).
The guide has been improved, mainly the usage section.
-------------------------- xfs211b2 ----------------------------
I've completely re-written the sector access / allocation on the
Amiga handler, so as to allow you to write to media larger than
floppy disks. Unfortunately I haven't had time to test it, so I
don't recommend you use it for anything crucial yet, just media
you can afford to re-format. Oh, and I found a few things were
missing in the directory cacheing (DC) part of the handler, so
I've made DC disks read-only for the moment till I sort it out.
Also, for some reason the utf-8 internal names were being sent
to the Amiga name hashing routines, causing all files whose names
contained accented characters to be put on the wrong hash chain.
If you have any disks written using this I suggest you use the
old version of the handler to copy the files off the disks then
use 211b2 or later to write them back (if it works!).
No further additions to any other handlers, 'cause I've been
adapting various compression routines, to see how easy it would
be to add archive-handling to xfsd.
I might as well mention that I had hoped to replace the rather
awkward DMS uncrunching in file.device, but I couldn't get access
to the source on the Aminet. Oh, didn't you know file.device could
read DMS files? Well I can't recommend it. Or DMS for that matter.
But you will need both, as file.device calls DMS to do the work.
-------------- xfs211b (only released on cover cd) --------------
MAC handler does something at last. You can now do directory
listings and read the data forks of files.
Temporarily disabled ExamineAll support so that ExamineAll should
now work properly. Yes you did read that right. If a handler
doesn't support ExamineAll the Dos emulates it via Examine and
ExamineNext. I'll re-enable it when it's fixed.
On the QL, accented characters should now be completely usable
in filenames, previously they would only read correctly.
Oh, and I've upgraded the x-code utility (to V1.01), and I've
replaced the geometry program with a more general device command
utility called devio (no docs for it but it's simple to use).
Oh, there seems to be some confusion as to the names of the
mountlists. I've made sure they're all named in the same way
(mount.#?).
Finally, I've changed all the flags fields in the mountlists
to make all volumes appear on the Workbench. So if you use
xfsd at the same time as another handler for the same disk,
something strange may happen (unless you edit the flags field
to stop xfsd mounting volume entries - see mount.xfsd).