JRD@USU.BITNET (Joe Doupnik) (01/24/89)
To correct a somewhat serious bug in MS-Kermit 2.32 (see next message), a
maintenance release of MS-Kermit -- 2.32/A -- is now available. The
opportunity was also used to fix a few minor problems. The changes from MS
Kermit 2.32 to 2.32/A 21 Jan 1989 are listed below.
Items common to all machines supported by MS Kermit -
1. Correct a problem negotiating 8-bit prefixing when MS Kermit is operating
with parity of NONE yet the other side requests that such quoting be done.
Version 2.32 had a problem and would receive the 8-bit quoting character
"&" as ordinary text and hence produce a corrupted file. Setting parity
to SPACE on the PC with version 2.32 might be a suitable workaround.
2. Correct ENABLE/DISABLE MESSAGE commands to work properly. Previously the
sense was inverted and the command applied to SEND rather than MESSAGE.
3. Formally decode messages in ACKs to data or filename packets before
displaying them.
4. Add analysis of prompt text in the ASK command so that \number forms are
converted to binary before displaying the prompt. This makes the prompt
field behave the same as in the ordinary SET PROMPT command. A dollar sign
will terminate the text string for either command.
Example: ASK \%1 \27[31mEnter password\27[37m
5. Correct small textual errors in the SET command main help screen.
6. Allow command REMOTE MESSAGE to use three character text messages.
Items specific to the IBM-PC version of MS Kermit -
7. Delay accessing a communications port until the port is actually needed.
This prevents the automatic probing of COM1 at Kermit startup, even before
file mskermit.ini or equivalent is read. COM1 remains the default port
unless another is selected by a SET PORT command. The effect of this change
is transparent except where Kermit has been operated under a multi-tasking
Environment such as Concurrent-DOS, Desqview, DoubleDos, or MS Windows.
These situations have experienced difficulty when a second copy of Kermit
is started while the first copy is actively using COM1; the disturbance
is created when COM1 is prepared for use as part of the Kermit startup
process. The present change prevents this disturbance by selecting but not
touching a port until a Kermit command requires it. Note that STATUS, SHOW
COMMUNICATIONS, SHOW MODEM, SET BAUD, SET SPEED, and SET PORT commands
do access the selected port for information; to avoid possible interference
between users of COM1 perform the desired SET PORT command to select and
initialize another port before using these latter commands.
8. Add several small improvements to video screen support for the VT102
terminal emulator when Video 7 boards are used. Preserve 43 line mode
whenever possible. Defeat 132 column mode for Video 7 VGA boards used with
fixed frequency analogue displays (31.5KHz), such as the IBM 85xx series.
Prevent using wrong display buffer address when an unknown proprietary
video mode is encountered from many EGA and VGA style boards.
9. Put the VT102 screen into the rollback buffer when clearing the entire
screen with any escape sequence combination. Partial screen erasures
do not cause preservation in the rollback buffer.
10. Support character 9Bh, ANSI "CSI", as equivalent to reception of "ESC ["
in the VT102 emulator.
11. Preserve character set pointers and related items in the VT102 emulator
when the screen width is changed.
12. Avoid unnecessary flow control commands when Control-Print Screen is
active.
The files which have been changed since version 2.32 are:
new release ident and date:
MSSDEF.H
new system independet files:
MSSKER.ASM, MSSSEN.ASM, MSSSER.ASM, MSSSET.ASM, MSSTER.ASM
and for IBM-PC's:
MSXIBM.ASM, MSYIBM.ASM, MSZIBM.ASM
Joe Doupnik
22 Jan 1989
[Ed. - Many thanks once again, Joe. The new files have been put in Kermit
distribution, and MSKERM.BWR has been updated accordingly.]
---
Date: Thu, 19 Jan 1989 18:42:32 CET
From: Helmut Waelder <ZRWA001@DTUZDV1.BITNET>
Subject: MS-Kermit 2.32 Binary Transfer Bug
I have transferred a binary file from IBM/370 Kermit 4.0 in server modus
to MS-Kermit 2.31 and a second time to MS-Kermit 2.32 with LOG PACKETS.
The result is as following (all data packets are the same for V2.31 and
V2.32):
------ MS-Kermit 2.31 ( works ok) ----------
Spack: ^A0 I~- @-#Y1~* ~G^M
Rpack: ^A- Y~% @-#&1~(Y^M
Spack: ^A/ Rdos33pat.arcY^M
Rpack: ^A- S~% @-#&1~(S^M
Spack: ^A0 Y~- @-#&1~* ~$^M <---- this packet differs in V2.32
Rpack: ^A/!FDOS33PAT.ARC.^M
-------- MS-Kermit 2.32 ( in error ) --------
Spack: ^A0 I~- @-#Y1~* ~G^M
Rpack: ^A- Y~% @-#&1~(Y^M
Spack: ^A/ Rdos33pat.arcY^M
Rpack: ^A- S~% @-#&1~(S^M
Spack: ^A0 Y~- @-#Y1~* ~W^M <----- different
Rpack: ^A/!FDOS33PAT.ARC.^M
Over the line comes the same data, but it seems to me that MS-Kermit does not
correctly interpret the 8th bit prefixing. There is no byte with 8th bit on
in the received file (v2.32) and the file is about 30% bigger in size.
/Helmut
[Ed. - This is a serious bug, and it is the main reason for releasing 2.32/A
at this time. If Kermit should be able to do anything right, it's file
transfer!]
---
The following files are MS-Kermit 2.32/A obtained from Columbia
and are now available via standard anonymous FTP from Simtel20:
pd1:<msdos.kermit>MSVIBM.ARC IBM-compatable specific
pd1:<msdos.kermit>MSVGEN.ARC Generic version
pd1:<msdos.kermit>MSVRB1.ARC Rainbow specific
pd1:<msdos.kermit>MSVHP1.ARC Hewlett Packard specific
pd1:<msdos.kermit>MSKERDOC.ARC User's guide and other files
pd1:<msdos.kermit>MSIIBM.ARC Script files, tutorial, init filesjkg@gatech.edu (Jim Greenlee) (01/30/89)
In article <KPETERSEN.12466619010.BABYL@WSMR-SIMTEL20.ARMY.MIL> JRD@USU.BITNET (Joe Doupnik) writes: >To correct a somewhat serious bug in MS-Kermit 2.32 (see next message), a >maintenance release of MS-Kermit -- 2.32/A -- is now available. The >opportunity was also used to fix a few minor problems. The changes from MS >Kermit 2.32 to 2.32/A 21 Jan 1989 are listed below. Before I get tons of mail on this, I want to let everybody know that I have a copy of 2.32/A. This is the version that I plan on posting in the near future unless I get inundated with requests not to. The consensus so far seems to be that posting Kermit would be a Good Thing, so I plan on doing it in the near future. Most folks seem to be leaning towards packaging the executables and docs separately. I should probably point out that Kermit is available via anonymous FTP from cunix.cc.columbia.edu (128.59.40.130 on Internet). The good stuff is under the kermit/a directory. I realize that not everybody can FTP, but if you're in a real hurry for it, that's one way to get it. It's probably a lot more reliable than netnews, too :-). Jim Greenlee -- Jim Greenlee - Instructor, School of ICS, Georgia Tech jkg@gatech.edu Cato said, "I had rather men should ask why my statue is not set up, than why it is." - Plutarch