home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Amiga Magazin: Amiga-CD 1996 November & December
/
Amiga-CD 1996 #11-12.iso
/
pd-disketten
/
ungepackt
/
8_95
/
apd-8-95-2
/
psion
/
docs
/
history
< prev
next >
Wrap
Text File
|
1996-01-22
|
4KB
|
110 lines
AmigaNCP-FileServer 1.6
-----------------------
- reversed "HideIcons" to be "ShowIcons", hiding .info and .backdrop
being the new default setting.
AmigaNCP-FileServer 1.7
-----------------------
- fixed to a misunderstanding on RF_OPEN with P_FCREATE / P_FREPLACE:
The path of the file wasn't created if it didn't exists. This caused
problems with Backup option of the S3a which relies on this behavior
and doesn't makedir() it's file's path.
AmigaNCP-FileServer 1.8
-----------------------
- changed to use "amigancp.library"
- fixed a bug in write handling where the last two bytes of a full
sized packet got trashed. Since internal data structures got
overwritten, this may have caused undefined behaviour or even
total system crashes.
- now correctly converts filenames from IBM to ISO charset.
- RF_SFSTAT (setting the file flags) didn't convert the name and
therefore always failed with E_FILE_NXIST. This is the cause of
error of the "File doesn't exists" requester after copying a file
from the S3's system screen.
==========================================================================
AmigaNCP-FileSystem 1.5
-----------------------
- fixed a bug in the file name conversion where the extension was cutted
at the first dot, not at the last. Both will fail, but this way
you are protected against side effects like "delete Agenda.Agn.info"
deleting "Agenda.agn" instead of simple failure.
- fixed a bug in ACTION_DELETE where invalid file names (e.g. containing
multiple extensions) to confusingly cause ERROR_DIRECTORY_NOT_EMPTY.
Now returns ERROR_OBJECT_NOT_FOUND.
AmigaNCP-FileSystem 1.6
-----------------------
- changed to use "amigancp.library"
- locks now contain a non-zero value in fl_Key. This should fix several
problems with programs making tricky assumptions about the internal
lock structure. <sigh>
- now correctly translates non-ascii characters in filenames from
ISO to IBM charset.
- now attempts to recreate the NCP channel if the remote NCP
changed (e.g. was killed and started again).
- fixed a race condition when Examine()ing the root dir: the
filename was actually filled in after the packet was replied.
- fixed a bug which caused random memory trashing if the first
entry in a directory scan was volume file.
- returned TRUE instead of DOSTRUE thoroughly.
- Examine()/ExNext() now returns the same value in
fib_DirEntryType and the (undocumented) fib_EntryType,
as the standard file system does. Should fix several buggy
directory utilities ;-)
- ACTION_DELETE always returned ERROR_DIRECTORY_NOT_EMPTY on
an error no matter what really happened.
- now returns ERROR_DEVICE_NOT_MOUNTED if the NCP link fails
for some reason.
- ACTION_EXAMINE on the root directory now correctly
returns the volume name.
- Locking "/" (or any combined form) didn't higher than the
pseudo subdirectories in NCP:.
==========================================================================
ncp.lib 1.1
-----------
- modified constructor to open "amigancp.library"
- fixed error requester to correctly show up the tasks name on
workbench startup.
==========================================================================
AmigaNCP-Monitor 1.3
--------------------
- relinked with ncp.lib 1.1 to use "amigancp.library"
==========================================================================
amigancp.library 2.0 (formerly called "ncp.library")
----------------------------------------------------
- "ncp.library" is now named "amigancp.library" in order to avoid problems
with Oxxi's Novell Client software, which uses a similary named
library.
- fixed several internal problems with reconnecting after the remote
side powered off without killing the actual NCP connection. This
always happened when the S3/S3a auto-powered off. Should now
correctly handle all cases of reconnection including such
nice operations as removing the link cable, turning off the
S3 link, reconnecting the cable and then turning back on
the link process.
- usage of the SERF_7WIRE protocol prevented llmac to "ping"
the S3 back into life after a power-off because DSR was low.
Now, if DSR is low, correctly closes the device, reopens it
without SERF_7WIRE to ping the S3 and then continues
normal operation.
- partial packets weren't properly discarded after a
reconnection.
- LLMAC didn't check data block sequence numbers. Now
data blocks which are not in sequence are silently
thrown away.