home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
EnigmA Amiga Run 1999 March
/
EnigmA AMIGA RUN 35 (1999)(G.R. Edizioni)(IT)[!][issue 1999-03].iso
/
earcd
/
-archivi
/
-recent2
/
ioblixz2usr378.lha
/
Changes
next >
Wrap
Text File
|
1999-03-17
|
4KB
|
145 lines
ioblixser.device
37.8:
- CMD_RESET could hang forever because of waiting for wrong allocated signals.
Fixed. Thanks to Albrecht Kinzkofer for finding this bug.
- sometimes it could happen that no more data could be sent due to a bug in
the interrupt handling code. Fixed.
- now DSR signal is checked on every write to avoid hangs when no DSR signal
is available and 7-wire mode is selected. Especially Term should now quit
without problems.
37.7:
- improved sending of data, system load reduced by ~10%. Before it was a
little bit faster, but asynchronous and completely uncontrolable.
- fixed the serial/parallel exclusion problem by using an own interrupt
chain
37.6:
- fixed a really bad bug when setting new parameters. Many many thanks to
Mathias Supp for helping to find this bug!
- some minor speed optimizations
37.5:
- fixed "hang up" bug during SDCMD_SETPARAMS
- now uses ioblix.resource as global database
37.4:
- added some safety checks during OpenDevice()
37.3:
- fixed bug in when reading in EOF mode
37.2:
- fixed carrier detection
- fixed Enforcer hits during OpenDevice()
- fixed Enforcer hits during CloseDevice()
- reduced system load by 50% during reading
- implemented ASDG-like SIOCMD_SETCTRLLINES command to control RTS and DTR
lines directly
ioblixpar.device
37.8:
- CMD_READ now reads data in a IEEE1284 compliant way
37.7:
- fixed the serial/parallel exclusion problem by using an own interrupt
chain
37.6:
- fixed a bug with wrong QuickIO handling
- now waits for SELECT signal from printer to avoid unwanted loss of data
- many thing have been simplified
- ECP status read back is much more stable now.
- ECP status is now read at every AllocECPInfo() call, not just when opening
the device
37.5:
- now uses ioblix.resource as global database
- PDCMD_QUERY returned wrong state of BUSY line. Fixed.
- ECP mode is now activated automatically if device is IEEE1284 compatible
37.4:
- added some safety checks during OpenDevice()
37.3:
- minor changes
37.2:
- fixed Enforcer hits during OpenDevice()
- fixed Enforcer hits during CloseDevice()
- hopefully fixed the "end of page not printed" bug
- fixed crashes during CloseDevice() for non-existing ports
ioblixzip.device
37.8:
- just a revision bump
37.7:
- just a revision bump
37.6:
- just a revision bump
37.5:
- now uses ioblix.resource as global database
37.4:
- added some safety checks during OpenDevice()
- some signals were not freed if initialization of Zip drive didn't succeed
- now supports normal parallel Zip drive and SCSI/parallel Zip Plus drive.
I hope it's ok, I only had two hours to make it work :)
IOBlix-Handler
37.8:
- just a revision bump
37.7:
- just a revision bump
37.6:
- just a revision bump
37.5:
- just a revision bump
37.4:
- added some small safety checks
37.3:
- internally set the startup string back to NULL to avoid problem with
bad programs
37.2:
- now the mount file's startup string is used to determine the device and unit
to be accessed
SetupIOBlix
37.8:
- UARTs oscillator frequency is now calculated for every board. This makes it
possible to use an 22.1184MHz oscillator to use 230400 baud and 480600 baud.
- using the option 22MHZ skips the frequency test and forces a frequency of
22.148MHz instead of 24MHz. This option is only necessary if you changed
your boards osciallor from 24MHz to 22.1184MHz.
37.7:
- fixed the serial/parallel exclusion problem by using an own interrupt
chain
- ioblix.resource has two new functions: AddIRQHook() and RemIRQHook()
37.6:
- fixed an Enforcer-Hit in AllocChipList() and FindChip() when searching for
non-existing chips
- fixed a really bad bug in ioblix.resource/AllocChipList() that made the
machine crash all the time
- some printers could be left in a state where switching off-line was
impossible. Fixed.
- UARTs FIFO size is now calculated by using the internal loop-back feature
instead of using a static hardcoded size.