Columbia Kermit/kermit.zip/b/krtupd.doc

Text File | 2020-01-01 | 8KB | 183 lines

KRTUPD.DOC   What's New in KRT and KRTTSX         V03.63       27-Sep-1997
----------   ------ --- -- --- --- ------         ------       -----------

This revision includes the following changes and updates since the
last formal (given to Columbia University) release -

1.  Added SET/SHOW CONTROL-CHARACTER ala C-Kermit.  If you can do it
there, you can do it here too, with the exact same command syntax.

This is useful for speeding up transfers, particularly binary data,
as it turns off control-char quoting, which uses two bytes to transmit
one non-printing char.  See HELP SET CONTROL-CHAR for details and usage.

Other Kermits that support this include MS-Kermit, its W95 successor
and C-Kermit.

2.  TSX EMT 114 is now used for output when TT is the link device
under TSX.  This makes Kermit _much_ zippier..  As the TSX docs say
ttyout loops are not handled very efficiently.

3.  Transmitted and received file attributes were made individually
settable and showable, ala C-Kermit.  These additions are documented
in HELP SET ATTRIBUTES and are useful when another Kermit has trouble
with some but not all of them.

4.  The file name conversion code was patched to accomodate various
previously troublesome Unix file names, such as .x. or .x: or ..x.
which were being erronously taken as implicitly wildcarded names.
Also all previously undefined chars in those possible in file names
were added, again for Unix compatibility.

5. Added SET FILE WILDCARDS - this doesn't modify RT-11 or TSX but
allows getting Unix files with a leading dot in the name while using
an output file to rename same (SET FILE WILD EXPLICIT to do this).
The default is IMPLICIT, as it always has been in K11/KRT.

6.  Added a warning message when receiving an empty file so one will
know why no data were transfered.

7.  Added a file name when opening LP as the log file so later
versions of RT-11 are happy.  As in the past LP must either be
spooled or extremely fast to serve as the log device.

8.  GETs to LP are now possible, although there's currently no
wildcard support when renaming files (whether to LP or a disk).

9.  CSISPC file size specifications, as in FILENAME[size], are now
supported and file sizes are now displayed when sending or receiving.

Note that for text files from foreign systems KRT adds 25% + 1 block
to the received size to accomodate differing line termination schemes,
and this increased size is what will be displayed, even though the
file will most likely be closed later at some smaller number of blocks.
If this bump up exceeds the currently available free space even though
the file 'should' fit a size which does may now be tried.

10.  Modifed the test for a Line Time Clock (KRTOSI) to accomodate
QED's fast machines, per a suggestion from Alan Sieving of QED.  It
now completes as fast as possible regardless of machine type, while
previously the QED CPUs finised the delay loop in less than one
clock tick.

11.  Added DAYTIME and SHOW DAYTIME commands.

12.  Use of a quoted string to place a blank at end of prompt string,
as in SET PROMPT "Kermit-11> ", is now supported.

13. Increased the command line input buffer to 132. bytes, in the EIS
assembly only.  This accomodates long command lines in take files and
of course also works for keyboard input.  Also fixed all possible sign
extension problems now that LN$MAX > 127.  As distributed KRTTSX.SAV
is built with EIS instructions and KRT.SAV is not.

14.  Added a T3000X.42 modem definition for XOFF restraint with this
modem.  Note that S48 must be set to 0 here or the modem will NOT do
XOFF restraint, regardless of what the modem docs say, at least in the
modem's ready to accept a command mode and sometimes when on-line to
another modem and exchanging data that need restraint.  This is a bug
in the modem itself and is the reason RTS/CTS flow control was added
to KM.  As long as Kermit is doing XOFF flow control (the only choice
possible other than with the KM handler) it will also be forced to
quote both XOFF/XON and their 8-bit counterparts, thus this won't
impact file transfers.

15.  An error packet is now sent to the other Kermit when appropriate
(as when a write to a local file fails) so the transfer gets properly
aborted.

16.  The file type (ASCII, binary, etc) is now displayed in the create
messages (to be sure you know what's going on) when transfering files.

17.  The contents of SEND FILE ACK and RECEIVE FILE ACK packets, such
as a file name returned by the remote Kermit, are now repeated char
decoded, 8-bit unquoted and then displayed during file transfers.

18.  The REMOTE command code now displays the reasons for retries as
each occurs.  Error packets are now sent when giving up, to attempt to
stop any possible lingering activity at the other end.  The BYE and
FINISH commands were moved into the REMOTE command processor so these
two now also use the above error handling enhancements.

Repeated char encoding is now done on REMOTE command arguments.  This
makes a big difference for a command like REMOTE DELETE *~ sent to a
Unix system, which without repeat encoding becomes 'rm *' after going
though C-Kermit and thus hoses everything.  8-bit quoting (for ASCII
chars > 177) is also now done (automatic when parity is in use) here.

19.  CD/CWD and REMOTE CD/CWD will now operate without an argument.
RT-11/TSX assigns are now tested for illegal string lengths as well.

20.  REMOTE WHO now passes arguments to the remote host.  Some hosts
support this, such as VMS where it's really handy on a busy system.

21. The first data packet is now actually rebuilt when resizing due
to its failing so the next retry is actually done with the smaller
sized packet.  In the past the reduction wasn't done until the next
packet was sent, which is of course too late..

22.  A rewind routine was added to KRTRMS.MAC for use when resizing
packets ala the above, instead of closing and reopening the file.

23.  An error in testing for terminal emulator command chars after
having SET CONSOLE 8-BIT is now fixed.

24.  The packet data processing routine (BUFFIL) was moved back to
the root for speed, as cleaning up made space available for it once
again.  This means less calling of overlays and helps quite a bit
when running on a floppies-only system.

25.  Added SPFUNs to the KM (Kermit Modem) comm handler to support
xmodem ala how it could be done with CL.  See KRTKM.MAC for details.

26.  Made "_" = "-" when parsing commands.  This is _only_ for KRT's
commands, any arguments and everything else remain untouched.

27.  Moved the HELP text index to a separate file (KRTIDX.MAC) and
added KRTIDX.BAS to create same, dumping KRTHLP.PRE and KRTHLP.BAS.
KRTHLP.MAC is now a permanent file.  KRTHLP had finally overflowed
4096. words, now there's plenty of room available.

28.  Modified the command definition macro (KRTCDF.MAC) to accept the
location of an optional prompt string instead of the string itself.
This allows sharing strings which reduced the overall program size by
2 blocks.

29.  Moved SHOW VERSION to KRTCMD and modified it so REMOTE HELP can
call and use it, thus displaying the exact same string one sees when
KRT is started.

30.  Fixed a  problem with the SET PHONE XMODE value specified in
a SET CL PORTS string from being overwritten by the default value
in the modem definition.  Now if a value is explicitly supplied it
will prevail.

31.  Moved all the pre-defined modem data to a separate module
(KRTMDM.MAC) to make room for more defintions there and to allow
for future expansion of the dialer module KRTDIA.  Moved SET$DTR
and SET$MODEM back to the dialer as there is now plenty of room
and this reduces the size of the unmapped (KRT.SAV) image.

32.  Fixed a problem with garbage being appended to a log file name.

33.  Fixed an unpopped stack on error exit from BUFEMP (receiving
a file).  This bug was introduced in V3.62 (but there was no error
checking at all before that), sorry!

34.  Any possibly lingering data from a modem is now flushed before
beginning dialing or redialing.

35.  All source files have been thoroughly cleaned up and fully
commented and the HELP text and doc files are completely caught up.

All program data have been moved to the tops of their source
modules to make foreign language translation as easy as possible.

The narrative edit history has been split out into its own file
(KRTHIS.DOC) so KRT.DOC is no longer the monster it was..

Billy Youdelman
[e-mail]
27-Sep-1997
-eof-