home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
8bitfiles.net/archives
/
archives.tar
/
archives
/
compuserve-file-archive
/
16 Vendor Support
/
XMODEM.THD
< prev
next >
Wrap
Text File
|
2019-04-13
|
72KB
|
1,930 lines
#: 14355 S1/Help/Data Lib Tools
16-May-88 17:44:02
Sb: #XMODEM - RunTerm
Fm: Bob Raikes 75616,1153
To: sysop (X)
I just spent a half-hour trying to download the top items in DL 6 and 8. I was
aborted immediately on each try. I use Runterm (stop laughing) and XModem
protocol. This worked for me a few months ago, the last time I could aafford to
use CompuServe. I can read the files, apparenlttly, but not download. Any idea
what the trouble could be? My bored son, desperate for a game, would be very
grateful.
#: 14357 S1/Help/Data Lib Tools
16-May-88 18:02:07
Sb: #14355-#XMODEM - RunTerm
Fm: Steve Nye/Sysop 76703,4032
To: Bob Raikes 75616,1153 (X)
Bob,
No one is laughing <smile>. We all have a personal preference.
In order to help you out, we need to know what happened. Did you receive
any error messages? Did the download even start to happen? Did you get a file
on your disk that you cannot use? Or did you get nothing at all? It may well
be that you got the programs, and don't realize it. If the files were .IMG
files (ending with
IMG) then when you use Xmodem you must use another program (offline) to remove
the Vidtex header. The program in DL1 called BINIMG.BIN will do this for you.
Let us know some of the details of what happened and we will try to help you
out. Runterm supports Xmodem, so there is no reason it shouldn't work.
SteveN
#: 14429 S1/Help/Data Lib Tools
22-May-88 18:34:50
Sb: #14357-#XMODEM - RunTerm
Fm: Bob Raikes 75616,1153
To: Steve Nye/Sysop 76703,4032 (X)
Thanks for your reply ,Steve. I was too frustrated with my first message to
give you the detail I should have. What is happening is that I am getting a
message saying operation aborted immmediately after runterm initiates the
download. To be honest, I don't know if it is a program-generated or CompuServe
generated message. The program always creates the file on disk before actually
starting the download. So what it does is close the file, leaving me with a
one-block file. The IMG factor is no problem. My version of runterm strips the
additional characters/. As Isaid, izI had no problem with the same program and
same procedure a few montshs ago, getting working program.ss okay. Oooops[[ To
bee perfectly precise, I should say that on some attemts pts everything hangs
up. Runterm prints the word waiting alternated with tranferriing block
(whatever) during a normal download. On some attempts now it just repeats the
word waiting every 30 seconds or so. I should also note that I've justried
several copies of the program and all respond the same, so it's not a corrupted
program. Whatvever, it's really frustrating. Any help you can give me, even if
it's just pinning fault on the program or Compuserve, will help. (I was
wondering if the new CompuServe software is a t fault; that('s one thing that
has changed since the program worked for me.) Anyway, thanks again.=
#: 14434 S1/Help/Data Lib Tools
23-May-88 18:04:06
Sb: #14429-#XMODEM - RunTerm
Fm: Steve Nye/Sysop 76703,4032
To: Bob Raikes 75616,1153 (X)
Bob,
It may very well be that the new software is just different enough to give
Runterm problems. I know when I use ProComm (on my IBM) there are two Xmodem
protocols, one for CompuServe and one for everyone else. Another thought, are
you using TeleNet, TymNet or something like that? The packet switching that
those services use sometimes causes problems.
I am fairly sure that the error message you are getting is from RunTerm. All
CompuServe error messages start with %NW??????? so they are readily
recognizable.
The waiting that is being done when you start does not surprise me. Be sure
you don't give up too early. Xmodem is a co-ordinated effort between your
computer and CompuServe and it can take up to 100 seconds to get the signals
correct. So make sure you aren't giving up after 30 seconds or so. In this
regard, it worries me that the transfer aborts so soon after starting. I am
using Xmodem and have all along, so I don't think it is the new software that
is doing it. I wonder if the timing in RunTerm may be at fault.
Another thought, there are two Xmodems...CRC and Checksum. I know
CompuServe recognizes CRC but don't know about Checksum. If RunTerm is using
that....it may be the problem. I am hoping another sysop will read this <HEY!
Jake!> and jump in with an idea here.
Wish there was an easy answer to your problem. But I haven't run into this
one before, and RunTerm is still new enough that not many folks know any of the
details...or its 'quirks'.
SteveN
#: 14438 S1/Help/Data Lib Tools
23-May-88 19:37:50
Sb: #14434-XMODEM - RunTerm
Fm: Ed Flinn 73127,1476
To: Steve Nye/Sysop 76703,4032 (X)
Steve,
"Normal" and "relaxed" versions of Xmodem are fairly common in MSDOS
programs. The original specs for Xmodem included some pretty short timeout
intervals that just won't cut it on CIS or any other PSN. While I've never
used Runterm, I have used a few dozen 64 term pgms, and have yet to see one
that wouldn't work here. There is, of course, always a first time.
(I used to own a Volks 6480. This modem has started many people on the
road to building a varied collection of terminal software. <grin>)
I'm sure that CIS supports checksum as well as CRC Xmodem. CRC is an
enhancement to greatly reduce the possibility of undetected errors. I believe
that CIS will attempt to use CRC first, but if the correct response to the
initial timeout is not recieved, will then switch to checksum. That's the
standard way of supporting both Xmodems. <ed>
#: 14442 S1/Help/Data Lib Tools
23-May-88 21:29:51
Sb: #14429-#XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: Bob Raikes 75616,1153 (X)
Bob:
Steve mentioned the two flavors of Xmodem, Checksum and Cyclic Redundancy
Check ("CRC" for short), the old version and the new version. The new one is
more reliable, according to the people who calculate every possible thing that
could go wrong. Compuserve will attempt to use CRC and then it switches to the
old checksum version. I have a terminal program that will try one Xmodem and
then switch if it fails. That might be the problem, but I doubt it (if CIS
tries 2-3 times with the wrong type, that would take 20-30 seconds or more).
I suspect your problem is the "prime-time" slowdown. The whole system slows
down when lots of people are using it. Most users prefer the hours 6 pm to
midnight. The four time zones have an overlap: 9-12 Eastern is the same as 6-9
Pacific time. If you're on the East Coast, try before 9 or after midnight. On
the West Coast, try before 6 pm or after 9.
ToddH
#: 14444 S1/Help/Data Lib Tools
24-May-88 00:23:52
Sb: XMODEM - RunTerm
Fm: Thomas R. Fry 70317,140
To: SYSOP (X)
To SYSOP(x)I have performed downloads previously using X - MODEM protocol. I am
using the same modem and terminal software. Apparantly, all I receive now are
repeated Control-x's (ASCII 24). Could you please advise ? Last DOW
approximately Dec. '87.
Thanks.
#: 14445 S1/Help/Data Lib Tools
24-May-88 11:21:22
Sb: #14442-#XMODEM - RunTerm
Fm: Gary Farmaner 76703,3050
To: Todd Heimarck/Sysop 76703,3051 (X)
Todd,
These start signals are only applicable to uploads. In Xmodem, the receiver
always begins the transfer process, and thus controls which version of Xmodem
to use.
I did a test, a while back, of the CompuServe start signals. CompuServe will
try to start a CRC upload ("C" sent) six times , and then checksum three times
(NAK) before aborting the upload.
The six attempts at CRC take about a minute.
Gary
#: 14447 S1/Help/Data Lib Tools
24-May-88 16:50:50
Sb: #14445-#XMODEM - RunTerm
Fm: Marte Brengle 76703,4242
To: Gary Farmaner 76703,3050 (X)
What drives me out of my mind is when I'm trying to up/down something and I've
got a single-sided disk in my 1571. I haven't got the new ROMS and that darn
drive will sit there and whir and flash and click and gronk and generally dink
around for what seems like ages before it says "Oh! You asked me to DO
something!" and gets on with it...
--M--
#: 14453 S1/Help/Data Lib Tools
24-May-88 23:28:34
Sb: #XMODEM - RunTerm
Fm: Thomas R. Fry 70317,140
To: Bob Raikes/sysop(x) (X)
After catching up on my reading of the messages, perhaps I can help Bob Raikes.
Also, I noticed Jose (?) also was having downloading problems. I have a
HESmodem II and use Runterm (no chuckles) with no problems......well,... I had
many succesful X-modem downloads back in Dec. '87 and everythingwas O.K. Due to
buisness obligations, I was offline until a week ago. I have never been so
frustrated in my entire CIS lifetime. I still (!!!) after many tries can't get
a succesful X-modem DOW. Yes, the the program will abort immediately. Yes, the
file is created and properly closed (with a <cr> as the only byte). I generated
a Source file of Runterm to aid my investigation. CIS sends countless
control-x's (ascii 24) ...why, I don't know. Runterm interprets this as a host
computer "abort command" ( at $2d8e ). It will also interpret a control-c
(ascii-3) the same way. What it does look for to start properly, is a countless
series of 0's. When it sees a control-a (ascii-1), it will start with the
transfer. If any sysop could help me find out what CIS has done to "their"
internal protocol, I could write a patch routine. I even changed the wait loops
....that waited so long and tried so many times...... that CIS gave up on the
transfer !! Incidently, Runterm does send out NAK's (one to grab the host's
attention).
Tom Fry 70317,140
#: 14456 S1/Help/Data Lib Tools
25-May-88 06:06:58
Sb: #14447-#XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Marte Brengle 76703,4242 (X)
Dear Marte,
I don't have the C128 ROMs either and so I continue to use each side as a
one-sided diskette. In other words, I format each side rather than using a
double-sided diskette -- otherwise it is certainly a waste of time for me.
Take care.
Sincerely,
Betty Knight
#: 14457 S1/Help/Data Lib Tools
25-May-88 06:18:59
Sb: XMODEM - RunTerm
Fm: Thomas R. Fry 70317,140
To: [F] Bob Raikes 75616,1153 (X)
After catching up on my reading of the messages, perhaps I can help Bob Raikes.
Also, I noticed Jose (?) also was having downloading problems. I have a
HESmodem II and use Runterm (no chuckles) with no problems......well,... I had
many succesful X-modem downloads back in Dec. '87 and everythingwas O.K. Due to
buisness obligations, I was offline until a week ago. I have never been so
frustrated in my entire CIS lifetime. I still (!!!) after many tries can't get
a succesful X-modem DOW. Yes, the the program will abort immediately. Yes, the
file is created and properly closed (with a <cr> as the only byte). I generated
a Source file of Runterm to aid my investigation. CIS sends countless
control-x's (ascii 24) ...why, I don't know. Runterm interprets this as a host
computer "abort command" ( at $2d8e ). It will also interpret a control-c
(ascii-3) the same way. What it does look for to start properly, is a countless
series of 0's. When it sees a control-a (ascii-1), it will start with the
transfer. If any sysop could help me find out what CIS has done to "their"
internal protocol, I could write a patch routine. I even changed the wait loops
....that waited so long and tried so many times...... that CIS gave up on the
transfer !! Incidently, Runterm does send out NAK's (one to grab the host's
attention).
Tom Fry 70317,140
#: 14460 S1/Help/Data Lib Tools
25-May-88 07:06:47
Sb: #14453-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Thomas R. Fry 70317,140 (X)
Dear Thomas,
Thanks for your message about XMODEM downloading. I forwarded a copy of
it to both Bob and Jose for their reading.
I have also asked some CompuServe people for the information you need to
possibly make a "patch." I'll let you know what I find out and perhaps the
CBMART Sysops or some of our Members will leave you an answer.
Thanks again for your interest and take care.
Sincerely,
Sysop/Betty Knight
#: 14464 S1/Help/Data Lib Tools
25-May-88 12:55:45
Sb: #14456-#XMODEM - RunTerm
Fm: Marte Brengle 76703,4242
To: Sysop/Betty Knight 76703,4037 (X)
Most of my disks are now in double-sided 1571 format, but some are not, such as
my Sidplayer disks, since a lot of my music was composed with the "old" Sid
editor and I didn't feel like transferring everything to new disks when I got
the Enhanced editor. So when I went to upload my Sid files the other night,
the drive went through its hoochie-coochie routine eight or nine different
times (however many tunes I uploaded). Argh. Why the heck can't the stupid
drive just read that "single sided" byte and say "Ah, this is a single sided
disk, I won't even bother checking the other side" ??
On the other hand, how much for a set of new ROMs?
--M--
#: 14471 S1/Help/Data Lib Tools
25-May-88 23:35:17
Sb: #14453-XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: Thomas R. Fry 70317,140 (X)
Thomas:
I just wanted to point something out. The (X) in parentheses after
someone's name is a marker that that person has read the message. So "SYSOP
(X)" means the message has been read by a sysop. You don't need to include the
X in your messages.
ToddH
#: 14472 S1/Help/Data Lib Tools
25-May-88 23:39:48
Sb: #14464-#XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: Marte Brengle 76703,4242 (X)
Marte:
If you have the old ROMs and a single-sided disk, you can force the drive
into 1541 mode in two ways:
If the computer is in 64 mode, turn the drive off and then on.
If you're in 128 mode, send the command OPEN 15,8,15,"U0>M0" and then
CLOSE15. Either way, it won't do the hootchie-kootchie when fed a
single-sider.
ToddH
#: 14477 S1/Help/Data Lib Tools
26-May-88 08:57:44
Sb: #14434-#XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Steve Nye/Sysop 76703,4032 (X)
CIS's XMODEM accepts either the CRC or Checksum operation.
I'd like to hear from the implementors of RunTerm. How can I get to them?
By the way, I'm the dude who maintains the protocol stuff for CIS, and to my
knowledge nothing has changed with XMODEM which should cause any such failures.
What got me was Thomas Fry's message saying that RunTerm "looks for a countless
series of 0's." That`s not teh way it says in the XMODEM document from Ward
Christenson.
On a download, XMODEM waits for the receiver to send a NAK (or C to indicate
CRC). Should this happen to get garbled on the way, it will try this three
times, waiting 20 seconds each time. If it dosen't get a NAK or C in that
time, it simply gives up.
Mr. Fry says that "CIS sends countless control-x's (ascii 24) ..." To my
knowledge, this should happen only if it first receives a CAN. The sequences:
<CAN><CAN>
<CAN><' '>
<CAN><CR>
or <CAN><timeout>
will result in a "cancellation", and result in the host transmitting a series
of three <CAN>'s in a row. Once the transfer has started, a timeout is
interpreted as a failure (XMODEM specs), and results in a single <CAN> being
transmitte.
It seems that almost eveyone who has implemented XMODEM has thrown in their own
personal twist, making it a difficult protocol to keep up with.
-Russ
#: 14478 S1/Help/Data Lib Tools
26-May-88 09:03:18
Sb: #14442-#XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Todd Heimarck/Sysop 76703,3051 (X)
Todd,
XMODEM is a "receiver driven" protocol. On a download, the sender waits for
the receiver to send a NAK, or an ASCII 'C' for CRC. Then the first packet is
sent.
For uploads, the receiver (that's CIS in this case) sends a C to initiate CRC
mode. The receiver will try sending the C several times (three in the case of
CIS), then try a few times with NAK, in the meantime waiting for the first
packet to arrive (the <SOH>).
I don't know what RunTerm is doing, but either it is not doing what is
expected or its timeouts are too short as you suggest.
-Russ
#: 14482 S1/Help/Data Lib Tools
26-May-88 13:41:24
Sb: #14447-XMODEM - RunTerm
Fm: Transactor HQ 76703,4243
To: Marte Brengle 76703,4242 (X)
Hi Marte:
I've got the old ROMs too and can sympathize. Just getting a dir seems to give
the thing conniptions. My cartridge helps though if I set single-sided mode.
Malcolm
#: 14483 S1/Help/Data Lib Tools
26-May-88 13:49:21
Sb: #14442-XMODEM - RunTerm
Fm: Gary Farmaner 76703,3050
To: Todd Heimarck/Sysop 76703,3051 (X)
I put those numbers backwards. It's "C" to start CRC three times, plus NAK 6
times for checksum, then abort.
#: 14484 S1/Help/Data Lib Tools
26-May-88 14:00:15
Sb: #14472-XMODEM - RunTerm
Fm: Marte Brengle 76703,4242
To: Todd Heimarck/Sysop 76703,3051 (X)
Todd, I know about those commands, thanks. I just haven't figured out how to
get them to work before I upload or download something with Sixth Sense. I
don't think I can send that command through the program, and since the disk
autoboots, the drive is in 1571 mode.
--M--
#: 14492 S1/Help/Data Lib Tools
26-May-88 18:56:43
Sb: #14477-#XMODEM - RunTerm
Fm: Steve Nye/Sysop 76703,4032
To: Russ Ranshaw (CIS) 70000,1010 (X)
Russ,
Glad to have the expert's opinion <smile>. I agree concerning Xmodem, it is
a hard protocol to keep up with...and now WXmodem, YXmodem, ZXmodem, and
God-only-knows-Xmodem!
I didn't really think that anything was put in the new software to change
the protocol, but several users have reported the same thing... that they were
able to use the same hard/software combination to download prior to the
software change...but not since. If it was only one or two, I would have opted
for operator error or a crashed disk or something else local. But there have
been sever (more than two) so the only common area I could think of was the sig
software.
Besides...it accomplished its purpose...it got the expert to drop by
<smile>.
You can get hold of Run Magazine at the following address:
RUN
80 Elm St.
Peterborough, NH 03458
They also operate a BBS....since it is work related <hehehe>....
(603) 924-9704
Thanks for dropping in and helping out.
SteveN
#: 14497 S1/Help/Data Lib Tools
26-May-88 21:53:21
Sb: #14478-#XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: Russ Ranshaw (CIS) 70000,1010 (X)
Russ:
A few comments about your messages to Steve and me regarding Xmodem and the
new CIS software:
First of all, the problem most likely is in the Runterm program and not in
your CIS routines. Or it might be in the hardware.
I use a program called "Bobsterm" and I've had no problems with Xmodem
downloads from CIS.
Second, there's a double-meaning to "zero." If you ask a Commodore modem
for a byte and there's nothing there, a null (no byte) can be interpreted as an
ASCII 0. The ASCII value of no-byte is sometimes an ASCII zero. When someone
says "the terminal waits for a series of zeros," he might mean "the term
program watches for something to happen."
Third, modem communications cause a non-maskable interrupt (NMI), and the
received byte automatically goes into the RS-232 buffer, which is only 256
bytes long. If the term software is written in interpreted BASIC, it might get
so far behind the buffer that the buffer overflows. Assembly routines never
have this problem at 300 or 1200. BASIC programs do.
Fourth, the RUNterm program was published in a magazine as a type-in
program. It's possible that the person complaining made a typing mistake,
although it's unlikely. Most Commodore magazines use checksum programs that
check your typing.
There's more, but you get the drift. I think the problem is in the RUNterm
program, not in the CIS software.
ToddH
#: 14498 S1/Help/Data Lib Tools
27-May-88 06:11:53
Sb: #14492-#XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Steve Nye/Sysop 76703,4032 (X)
Steve,
Hmmm. I had thought someone here could give me the source. I personally
don't have time to call RUN magazine or their BBS.
I agree that it seems the common items here are our XMODEM software and the
RunTerm program. I remember 'way back when, when someone would submit a
FORTRAN deck to the university compter and complain that it didn't work, that
it worked six months ago, that the submittor hadn't changed anything, that it
simply _HAD_ to be the computer or the compiler or the operating system or
whatever. And almost _every_ time it turned out that the dude had in fact made
a minor change, and tried to cover it up by saying "Oh, that little change
shouldn't make any difference."
I see no mention of port parameters in teh messages or replies thereto.
XMODEM, and most protocols except Kermit, require 8 databits, no parity. Has
this issue in fact been pursued, or does it not make sense for Commodore
computers (of which I am nearly ignorant)?
Is the user doing whatever is necessary to initiate the XMODEM receive on his
end?
By the way, the last time the XMODEM code was touched was February 1, 1988,
and that was to change error codes returned to the calling program. No changes
were made to the logic. So, I'm at a loss here. The current XMODEM code has
been in use since then, and these are the first reports I've seen, nearly four
months later.
Can the more knowledgeable in the Commodore world pursue this a little more
thoroughly, since I don't have the necessary credentials?
-Russ
#: 14499 S1/Help/Data Lib Tools
27-May-88 06:12:01
Sb: #14497-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Todd Heimarck/Sysop 76703,3051 (X)
Todd,
Ah, I forgot about the <NUL> and its role in Commodore systems. Dosen't that
make XMODEM rather difficult to implement since it has no support for quoting
<NUL> or anything else? It is possible to get a checksum of <NUL>, so I don't
really see how it can work.
Since XMODEM uses a fixed packet of 132 bytes, I don't see this as a problem
for the 256 byte comm. buffer.
I'll drop out of this, then, until some more definitive data comes to light.
-Russ
#: 14500 S1/Help/Data Lib Tools
27-May-88 08:03:39
Sb: #14498-#XMODEM - RunTerm
Fm: Bob Raikes 75616,1153
To: Russ Ranshaw (CIS) 70000,1010 (X)
Gee, guys, what a broad range of opinions. Okay, let's get one thing clear:
There was NO change to my program, and probably not to that of the other users
who complained. Not a little change, No change. Also, as I said, I tried
several versions, so it wasn't a corrupted disk. As I said, downloads with
Runterm before the first of the year were consistently successful. I was off
for months, came back in mid-May, and found that I could not download. I'll bet
the house that the software changes are the cause. I'm not a programmer so I
don't follow completely the arguments presented in this case. Thomas Fry seems
to be willing to run this down; that would be very nice and appreciated. I
personally am an underpaid newspaper drudge and can't afford to buy another
terminal program, so either a patch or a CIS repair would help my youngsters
continue to learn their way around onlineland. Thanks for all the responses but
don't give up.
#: 14501 S1/Help/Data Lib Tools
27-May-88 10:39:47
Sb: #14498-#XMODEM - RunTerm
Fm: Marte Brengle 76703,4242
To: Russ Ranshaw (CIS) 70000,1010 (X)
I am SO glad to see you make mention of the fact that Xmodem requires certain
word length and parity settings. I just got a message on CBMCOM a day or so
ago in which someone berated me for saying just that, insisting that Xmodem
"doesn't care" about such things and simply downloads exactly what is sent to
it regardless of how the program is set. Perhaps you could come over to CBMCOM
and tell the gentleman yourself, since he apparently doesn't think I know
what's what <smile>
--M--
#: 14502 S1/Help/Data Lib Tools
27-May-88 11:47:45
Sb: #14500-#XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Bob Raikes 75616,1153 (X)
Bob,
My observations are based on what I have experienced over many years in the
saddle. I wasn't accusing anyone of willful misinformation, only pointing out
that there are many paths to traverse and scenic vistas on the way to
uncovering what is actually going on rather than taking pot shots at one end or
the other.
We have a lot of users, many of whom use XMODEM to do up- and down-loads with
the Data Libraries. When someone comes along with a tale of woe concerning
XMODEM, I have to begin questioning to find out exactly what is happening.
You indicated that you tried XMODEM downloads from CompuServe and they
failed. Did you try any from a local BBS?
I would like to have a step-by-step walkthru of what happens, from the time
that the download is requested on the host until you finally give up. I'd also
like to have some information, such as the timing parameters of the XMODEM
implementation in use, and anyother piece of evidence which you might be privy
to.
I'll bet your house that software changes on our end are not the cause of
your problems! :+) About the only thing I can think of at the moment is that
RunTerm has classic XMODEM timing, making its use on CoumpServe very iffy, or
the port parameters on the coumputer aren't 8N1.
-Russ
#: 14503 S1/Help/Data Lib Tools
27-May-88 11:47:55
Sb: #14501-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Marte Brengle 76703,4242 (X)
Marte,
You might point out that the second byte in an XMODEM packet is the sequenc
number, and that it has a range of 0 to 255, which requires 8 bits, and that
the checksum can also assume any value in the range of 0 to 255, and that
XMODEM is used to transfer binary (8 bit) as well as ASCII data. Oh, yes, he
might have to be told that in order to cope with the range 0 to 255 requires 8
bits, and that is the exact size of bytes in our computers and of the data
shipped over communication lines. You can cite me as your source! :+)
-Russ
#: 14506 S1/Help/Data Lib Tools
27-May-88 11:58:32
Sb: #14501-#XMODEM - RunTerm
Fm: Gary Farmaner 76703,3050
To: Marte Brengle 76703,4242 (X)
Marte,
It's possible that Tony is RIGHT.
CompuServe statements are based on the hardware/software they know. Hardware
UARTs when set to a specific parity case (except none) may strip the parity bit
before giving the byte to the computer. Therefore, if you don't convert to
8/1/NONE you'll lose the high bit.
What I beleive Tony is saying is, that the Commodore 8-bit UART software
routines DO NOT modify the byte received. A parity error will be flagged, but
the actual byte in the buffer still has the parity bit left untouched. Thus,
the byte is as if the system were set to 8/1.
I'll have to investigate.
Gary
#: 14507 S1/Help/Data Lib Tools
27-May-88 14:13:06
Sb: #14506-XMODEM - RunTerm
Fm: Marte Brengle 76703,4242
To: Gary Farmaner 76703,3050 (X)
Gary, Tony didn't say anything about "Xmodem on the Commodore." He made it a
GENERAL statement, and that is why I took issue with what he said (leaving
aside the tone in which he said it). The problem with making blanket
statements like that ANYWHERE is that they simply muddy the issue.
For whatever reason, Commodore computer owners ARE having trouble with Xmodem
downloads lately, and I feel it's worth changing whatever we can on our end
just to see if it makes a difference. If it doesn't, THEN we assume it's a
system problem. But if Tony is right and the settings on our computers don't
matter, then obviously something is massively wrong somewhere.
--M--
#: 14520 S1/Help/Data Lib Tools
28-May-88 19:11:22
Sb: #14501-#XMODEM - RunTerm
Fm: John F Davis 73455,43
To: Marte Brengle 76703,4242 (X)
You know, Some people think they know everything. Some DO know everything (at
least about some things)
X-MODEM does not care about the leangth of the words (text) it downloads. What
the KNOW-IT-ALL (read that as *** **** (word not postable)) fails to consider
is that WORD LENGTH means something quite different to a computer user then it
does to a typeist!
If you hear from him again refer him to me.
John
#: 14521 S1/Help/Data Lib Tools
28-May-88 21:04:49
Sb: #14502-#XMODEM - RunTerm
Fm: Bob Raikes 75616,1153
To: Russ Ranshaw (CIS) 70000,1010 (X)
Thanks a lot for offering to help figure this out. It's well over my head. The
suggestion about trying Runterm with another board is a good one, but,
unfortunately, I don't even know of any. I don't do this a whole lot. Hmm. I do
have a list of some; I'll try it and let you know how I make out. I don't know
about timing and other technical details of the program. I can try to upload it
-- that's something I haven't tried yet -- and I'll query the magazine. Or I
can otherwise make the code available if you can tell me how. As for the
disastrous details: There has been no change in hardware -- and the system
still runs Flight Simulator II and other complex stuff, so I think it's OK --
and I have about 4 copies of the program, all of which give the same results.
Before the first of the year the only time or two I had download failure was
when I got kicked off the line on a weathery night. Since I came back this
month what happens is that as soon as I hit the return key after telling
Runterm to download, either one of two things happens. I get an immediate
<operation aborted> message, or Runterm waits indefinitely -- I let it go 5 or
6 minutes -printing the word WAITING, which it should print at the time it
tells the host to transmit. This is at any time of day, and last December I
never had to wait more than 10 seconds or so and seldom that long. Runterm
still reads files OK, but that's no way to obtain machine language programs
unless you live for frustration. The program's operation is goof-proof, so it
can't be that I've forgotten how to use it -- and don't forget that there have
been others with the same problem. Well, I'll try uploading it. Thanks a lot
for your interest and offer to help.
#: 14526 S1/Help/Data Lib Tools
28-May-88 23:02:57
Sb: #14520-XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: John F Davis 73455,43
John:
Gary Farmaner pointed out on CBMCOM that for some computers, word length
**does** make a difference.
Some computers use a chip called a "Universal Asynchronous
Receiver/Transmitter" (UART) for connecting to modems. If you send the UART a
command to ignore the eighth bit, it throws it away. That means that the
Xmodem checksum has a 50-50 chance of being wrong, since the final bit is
always going to be zero and the checksum might or might not include a one
there.
Commodores don't use a hardware UART; it's all handled in the operating
system software. This means terminal programs pay attention to the eighth bit
during Xmodem transfers.
ToddH
#: 14540 S1/Help/Data Lib Tools
30-May-88 01:42:00
Sb: #XMODEM - RunTerm
Fm: Thomas R. Fry 70317,140
To: Bob Raikes 75616,1153 (X)
Hi Bob. I located the problem and I have the solution.
Runterm was at fault, but CIS was allowing it work until they changed;
probably last February.
-Load Runterm (but don't run it)
-In Direct Mode, Type In:
POKE 11510,0
and press [RETURN].
-SAVE a copy, or RUN it.
I just performed a download of DL1.DIR succesfully, and it seemed to work
fine. I did get a few block re-sends, probably
from noisy Telco lines.
If anyone is curious, I'ld be more than happy to explain the how's and
why's.
Happy Downloads !
Thomas Fry,70317,140
#: 14542 S1/Help/Data Lib Tools
30-May-88 12:15:46
Sb: #14540-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Thomas R. Fry 70317,140 (X)
Dear Thomas,
Your "RunTerm" poke 11510,0 corrects the error when downloading with its
XMODEM. Congratulations and that's terrific!!
I'm more than curious to know what was causing the XMODEM download Error.
I will forward your message on to Russ Ranshaw of CompuServe and your answer to
this message. You know that some Members besides Commodore users have also
noted a problem with XMODEM downloading and your patch here may be the answer
that CompuServe needs.
I'm sure that the other Sysops and Members will be looking forward to the
"how's and why's" of this XMODEM problem. Thanks and take care.
Sincerely,
Sysop/Betty Knight
#: 14546 S1/Help/Data Lib Tools
30-May-88 23:11:28
Sb: #14540-XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: Thomas R. Fry 70317,140 (X)
Thomas:
I'd be interested in an explanation, too. If it's a long one, you might
want to upload it.
Todd Heimarck
P.S. In fact, it might be a good idea to make it a permanent part of the
Help section. If I were you, I'd put the POKE in the descriptionof the file.
Then, if people wanted to read the hows and whys, they could download. If they
just wanted the POKE, they could read the description without downloading.
#: 14547 S1/Help/Data Lib Tools
31-May-88 07:19:28
Sb: #14521-#XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Bob Raikes 75616,1153 (X)
Bob,
I saw a message indicating how RunTerm can be "poked" into working. I have
no idea what the poke does; if anyone knows, PLEASE let me know!
Have you tried the fix?
-Russ
#: 14548 S1/Help/Data Lib Tools
31-May-88 10:17:43
Sb: #14547-#XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010 (X)
Dear Russ,
Yes, the "poke 11510,0" to the "RunTerm" Terminal Program certainly does
work for downloading with its XMODEM - I've tested it. We all are looking
forward to Bob Raikes' message about the explanation of what this poke does.
If the answer doesn't come to your PPN I will be sure to "mail" you a copy
through EasyPlex. Whatever it is - it may answer some XMODEM problems with
other brands of computers. Thanks for your interest; I appreciate it. Take
care.
Sincerely,
Sysop/Betty Knight
#: 14550 S1/Help/Data Lib Tools
31-May-88 17:01:58
Sb: #14548-#XMODEM - RunTerm
Fm: Ed Flinn 73127,1476
To: Sysop/Betty Knight 76703,4037 (X)
Betty,
Uh oh, I think there may be more than one version of Runterm, then. Over
the weekend I went and hunted down the version you mentioned, from SOFTEX. It
doesn't seem to have any mention of a particular version on any screen or
message I've seen. The POKE 11510,0 doesn't help that version any. In that
version, in fact, at 11510 is the first letter "o" in the "Upload or download"
message. Performing the POKE just causes that message to be truncated, since
the program, apparently, is delimiting messages with $00. The download
continues to fail. With any luck Bob Raikes' explanation will allow the
offending byte in this version to be found and fixed, as well. <ed>
#: 14554 S1/Help/Data Lib Tools
31-May-88 22:34:39
Sb: #14548-#XMODEM - RunTerm
Fm: Bob Raikes 75616,1153
To: Sysop/Betty Knight 76703,4037 (X)
My son called from home and said Thomas Fry's one-bite repair to RunTerm
worked. Amazing. I haven't the faintest idea what is in that part of memory
and, having decided a few years ago that I'd have to give up supporting my
family if I wanted to seriously study 6502 machine language, I probably
couldn't figure it out. The gentleman said he'd explain if people want to hear
an explanation. Obviously there are at least a handful of us who won't sleep
right until we know. I'd also like to know why the program worked for so long
and how the CIS software changes ended that cozy relationship. I'd very much
like to thank Mr. Fry.
#: 14558 S1/Help/Data Lib Tools
01-Jun-88 07:16:33
Sb: #14550-#XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Ed Flinn 73127,1476 (X)
Dear Ed,
Thanks for your message and you are absolutely correct about two versions
of "RunTerm."
The "RunTerm" patch "poke 11510,0" works with the version that Bob
uploaded privately for research purposes. I first downloaded "RunTerm" from
CompuServe's SOFTEX area and tried to download with its XMODEM and it wouldn't.
Then I tried with Bob's version and it would't download either.
Then Thomas Fry wrote about the "patch" and I tried it with Bob's version
and XMODEM downloaded fine. I did NOT test with the patch in the current
SOFTEX version.
Last night, after reading your message, I tried the "patch" in the SOFTEX
version and it doesn't work as you wrote. So there must be two versions. We'll
need to wait for Thomas' explanation of the "patch" so we can then find the
correct location in the new version. I'm realy curious about it and why what
used to be allowed in the old Access Software which isn't allowed now.
Thanks again and take care.
Sincerely,
Sysop/Betty Knight
P.S. These two versions have different file sizes also.
#: 14560 S1/Help/Data Lib Tools
01-Jun-88 08:35:18
Sb: #14554-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Bob Raikes 75616,1153 (X)
Dear Bob,
I'm glad that the "RunTerm" patch worked for your Son. Most of us are
waiting anxiously for Thomas Fry's message that explains what the patch does.
You might want to read messages #14550 from Ed Flinn and my answer #14558
about there being two versions of "RunTerm" around. Just enter at the Forum!
or Messages ! prompt -- ri 14550 and then read the reply to it.
Good luck and take care.
Sincerely,
Sysop/Betty Knight
#: 14567 S1/Help/Data Lib Tools
01-Jun-88 18:04:31
Sb: #14498-#XMODEM - RunTerm
Fm: Steve Nye/Sysop 76703,4032
To: Russ Ranshaw (CIS) 70000,1010 (X)
Russ,
I didn't mean to imply an error in the CIS code...but rather that some
change in the CIS code had made this Xmodem protocol MORE particular than it
had been in the past. Since several users had the same problem, and had in the
past used the software successfully, I was searching for something (like the
parity you pointed out) that might be more critical now than it had been with
the old software. But as someone has already found the fix (I wonder what is
at that address anyway <smile>) the point is moot (or it is mute....or just
unimportant <hehehe>)
Thanks again Russ.
SteveN
#: 14575 S1/Help/Data Lib Tools
01-Jun-88 23:00:52
Sb: #14558-XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: Sysop/Betty Knight 76703,4037 (X)
Betty:
I think we could find the byte to POKE in the other version. If I could
download both versions from one of the sections reserved for sysops, I could
probably locate the byte to POKE.
ToddH
#: 14578 S1/Help/Data Lib Tools
02-Jun-88 09:08:45
Sb: #14567-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Steve Nye/Sysop 76703,4032
Steve,
Having seen the explanation by Thomas Fry, it seems that RunTerm is doing
something that is not covered in the XMODEM specification, namely sending a
message after the NAK, which the host interprets as in-appropriate garbage and
decides that the link is corrupted so attempts to cancel the transfer.
Looking over the XMODEM code here, I don't understand why it should be
happening. I made changes to the logic to prevent such problems a while back,
and the production version of ACCESS might not yet have the upgrade in it.
Please try the current FILTRN on Level 5 to see if it causes the same kind of
problem using un-modified RunTerm.
To use the Level 5 FILTRN, enter the command
PER
(PERsonal area). You will be taken to the "OK" prompt. Enter the command
LEV 5
Then
R FILTRN
Select XMODEM protocol and answer the questions. If you don't have any files
in your personal area (check with the DIR command before doing the R FILTRN),
upload one first.
I think it should work. If not, let me know. I'll send you a copy of
our XMODEM code which is the likely spot for causing the reported problem.
Maybe someone else can spot something I can't.
-Russ
#: 14579 S1/Help/Data Lib Tools
02-Jun-88 09:09:07
Sb: #14567-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Steve Nye/Sysop 76703,4032
[GET_ACK] :
BEGIN
! Try to get a character
IF (CHAR = GET_X_CHAR ($TIME_ACK)) EQL FAILURE
THEN
ACTION = Error
ELSE
SELECTONE (.CHAR AND %O'377') OF
SET
! Got the <ACK>, successful transfer
[$CH_ACK] :
BEGIN
SEQUENCE = .SEQ_NEXT;
RETURN SUCCESS;
END;
! Negative acknowledge, resend buffer
[$CH_NAK] :
ACTION = ERROR; ! 8 bit nak
[%C'C'] :
ACTION = ERROR; ! 16 bit nak
! Receiver wants to cancel
[$CH_CAN] :
BEGIN
! Check if it is really a CANcel
IF (CHAR=GET_X_CHAR ($CAN_TIME)) EQL FAILURE
THEN
ACTION = CANCEL
ELSE
BEGIN
IF ((.CHAR AND %O'377') EQL %C' ') OR
((.CHAR AND %O'377') EQL $CH_CR) or
((.Char and %o'377') eql $ch_CAN)
THEN ACTION = CANCEL ! Some send CAN SPC
! others send CAN CR
! yet others send CAN CAN
ELSE ACTION = ERROR; ! Must be garbage
END;
END;
[OTHERWISE] :
begin
ERR_COUNT = .ERR_COUNT + 1; ! Increment error count
if .Err_Count gtr $Err_Max
then action = Cancel
else ACTION = GET_ACK;
end;
TES;
#: 14598 S1/Help/Data Lib Tools
03-Jun-88 11:52:48
Sb: #14578-#XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010 (X)
Dear Russ,
We found there are two versions of "RunTerm" and neither one would
work. The "patch" in the older version works fine - we don't have a
"patch" for newer version yet.
I tried with both versions in my PERsonal File Area with R FILTRN
to download with XMODEM. Neither one would work on both LEV 5 and LEV
1 (naturally NO on LEV 0 either).
Without the "patch" in either version on all LEVels I get the
following error messages: --
>>OPERATION ABORTED<< (From "RunTerm" - ?)
and then --
? FTRABT - Transfer Aborted! (I assume from CompuServe)
We're working on a "patch" for the newer SOFTEX version. Thanks
for your help and take care.
Sincerely,
Betty Knight
#: 14601 S1/Help/Data Lib Tools
04-Jun-88 12:17:02
Sb: #14598-#XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Sysop/Betty Knight 76703,4037 (X)
Betty,
Are you saying that neither unpatched RunTerm works with FILTRN from
any level? Most odd; I'd expect them to work with the new version on
Level 5, or 1 whichever it's on.
Evidently the >>OPEATION ABORTED<< is from RunTerm; nothing we have
puts out anything that looks like that. The ? FTRABT ... is from
FILTRN.
I'll see if it can be duplicated with anything I have sources to.
-Russ
#: 14603 S1/Help/Data Lib Tools
04-Jun-88 18:19:49
Sb: #14601-XMODEM - RunTerm
Fm: Steve Nye/Sysop 76703,4032
To: Russ Ranshaw (CIS) 70000,1010 (X)
Russ,
I THINK the problem can be duplicated like so....
Initiate the Xmodem download on the CIS end.
Before starting the download at the remote, send the following:
<CR> Text <CR>
The text is unimportant, and I think the second <CR> is unimportant.
The system sees the carriage return, and aborts. There is a prompt
that Runterm sends to the host at the beginning of the transfer, and
that prompt is bracketed by returns. I believe that CIS sees the
returns and aborts. At any rate, if you follow the above procedure,
the download aborts long before CIS has had the time to time-out
(approximately 100 seconds if I am correct )
So the problem is the text line that Runterm sends the host, but I
wonder why CIS used to ignore it, and now aborts.
SteveN
#: 14607 S1/Help/Data Lib Tools
05-Jun-88 08:41:54
Sb: #14601-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010 (X)
Dear Russ,
Yes, I am saying that both "RunTerm" C64 Terminal Programs will
not download with XMODEM with R FILTRN from my PRO area using LEVels
0, 1 and 5. This means that SOFTEX has a Version of "RunTerm" which
will not download.
CompuServe did have a C64 when I "beta tested" C64 VIDTEX for
Duane Harris, Randy Raynak and John Pampuch. If you still have it you
could see the problem using the SOFTEX's "RunTerm" program Number 909.
Thanks for your interest and take care.
Sincerely,
Betty Knight
#: 14658 S1/Help/Data Lib Tools
08-Jun-88 12:33:27
Sb: #14456-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Sysop/Betty Knight 76703,4037 (X)
Betty,
There is a new version of FILTRN on Level 5 which has some changes to
XMODEM support. Can you please try RunTerm and let me know what
happens?
For the edification of yourself and others, the situation is that
back in December we were fielding complaints from a commercial customer
that our XMODEM code as not behaving "properly" (what ever that
means!). In particular, we were retransmitting a packet whenever
spurious characters were received while waiting for the ACK on
downloads. The perticular software that the customer was using didn't
like duplicated packets, and would crash if too many duplicates were
seen. So, a change was made to not automatically retransmit on a
timeout or spurious characters (that is, not an ACK, NAK, or CAN), but
each spurious character was counted as an error. After ten errors
while waiting for an ACK, we would send a series of three CANs.
The change on Level 5 is to not count spurious characters as errors.
I might point out that a lot of this is not covered in any XMODEM
documentation I have ever read. In fact, most such documentation
assumes that there is a point-to-point communication connection between
the two machines, and that instantaneous response to a NAK etc is
possible. But such response is _NOT_ possible over a switched network
such as CompuServe's, and so changes had to be made to allow for the
nature of networks.
-Russ
PS: The updated XMODEM support will not be in production for several
weeks.
#: 14659 S1/Help/Data Lib Tools
08-Jun-88 13:14:47
Sb: #14658-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010 (X)
Dear Russ,
Thank you for your very informative message concerning XMODEM and
I will test LEV 5 of FILTRN with the "RunTerm" Terminal Program (both
versions of "RunTerm"). I'll do the test this afternoon and leave
you the results in an answer to your message. Thanks again and take
care.
Sincerely,
Betty Knight
#: 14668 S1/Help/Data Lib Tools
08-Jun-88 22:31:11
Sb: #xmodem - Runterm
Fm: Thomas R. Fry 70317,140
To: [F] Sysop/Betty Knight 76703,4037 (X)
Hi Betty. Sorry to create such a frenzy. I have not been able to keep
up with
all the messages, but I still say that Runterm is at fault. It
violates the prime protocol. Yes, it does print OPERATION ABORTED (I
thought I said that once). I hope that simply correcting the user
tel-com software would solve/end it.
Sorry,
Tom Fry 70317,140
#: 14674 S1/Help/Data Lib Tools
09-Jun-88 08:44:36
Sb: #14658-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010
Dear Russ,
I've spent quite a bit of time testing these two "RunTerm"
versions. I uploaded a small Binary program file to my PERsonal Area
so I could test with program files as well as ASCII files. I
downloaded my test .BIN file using "VT100-128" and the program
downloaded and ran just fine.
Using the older version of "RunTerm" with no patch I went to my
PER Area and set Level 5 at the R FILTRN prompt. The download did NOT
abort like it does on LEVel 0 but the program will NOT run when loaded.
So then I loaded the newer version of "RunTerm" from SOFTEX with NO
patches and repeated the above test. This time I received the
"RunTerm" message of "Transfer in progress" and then a series of ten
"RunTerm" prompts of "waiting" because CompuServe did not start their
transfer and "RunTerm" timed out after ten "waiting" messages with
">>OPERATION ABORTED<<"
Sysop Todd Heimarck did some research with the two versions of
"RunTerm" and he created a "Patch" for the SOFTEX version -- poke
7698,0
and Thomas Fry created one for the older version -- poke 11510,0
With each version with its noted patch above, they both download
fine from the CBMART LIBraries using XMODEM and also my test Binary
upload in my PERsonal Area.
The conclusion I came to from the testing is:
The LEVel 5 FILTRN did not work
Both programs with the "poke" patch do work all right
Does anyone have any opinions on this? Thanks and take care.
Sincerely,
Betty Knight
#: 14675 S1/Help/Data Lib Tools
09-Jun-88 09:06:11
Sb: #14668-##xmodem - Runterm
Fm: Sysop/Betty Knight 76703,4037
To: Thomas R. Fry 70317,140 (X)
Dear Tom,
I really appreciate your interest with this "RunTerm" XMODEM problem. Your
patch of "poke 11510,0" certainly corrected the problem with the older version.
Sysop Todd Heimarck discovered that the patch "poke 7698,0" corrected the
problem in the newer SOFTEX "RunTerm."
The one thing that bothers me is -- why did these versions work before? I
suppose we will never know but we do now have patches for both versions. Thanks
again, Tom, for your help and interest in this problem. Take care.
Sincerely,
Sysop/Betty Knight
#: 14682 S1/Help/Data Lib Tools
09-Jun-88 16:50:17
Sb: #14674-#XMODEM - RunTerm
Fm: Ed Flinn 73127,1476
To: Sysop/Betty Knight 76703,4037 (X)
Betty,
I think there's a problem now in LEV 5. In order to pursue this latest
wrinkle, I used a disk doctor to create a one sector file of alternating bytes
of $00 and $FF. At the default level in R FILTRN, I uploaded this as binary,
using DarkTerm, and immediately downloaded it with DarkTerm, and checked it.
All OK. I then logged on with Runterm, issued LEV 5, and downloaded it. All
$00's were stripped from the file on download. Since this program has
apparently worked in the past, I logged back on with DarkTerm, issued LEV 5,
and downloaded the file. Again, all $00's were stripped from the file. I
believe the LEV 5 Xmodem is not properly handling nulls. <ed>
#: 14696 S1/Help/Data Lib Tools
10-Jun-88 07:10:57
Sb: #14674-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Sysop/Betty Knight 76703,4037 (X)
Betty,
Please do a VER command to FILTRN and tell me what version is running. It
should be (113) or later. Also, try this:
o R FILTRN
o Select XMODEM
o Etc.
o Instead of starting XMODEM on your end,
enter control-U (a NAK).
o FILTRN should send the first packet.
o Type something like OK, SEND THE FILE, hit enter
(or return or whatever it`s called).
Nothing should happen. If you get the control-X's
(I have no idea what they do on a Commodore),
it is probably the wrong version of FILTRN.
This is the test I used to verify that the reported problem for RunTerm was
fixed. Am I missing something?
-Russ
#: 14697 S1/Help/Data Lib Tools
10-Jun-88 07:11:05
Sb: #14682-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Ed Flinn 73127,1476
Ed,
Did you select "Binary" when you did the download? ASCII mode will strip
NULs, but Binary dosen't (or shouldn't). Other people regularly download
binary files with no ill effects on a variety of systems. Oh, dosen't the
Commodore have problems with <NUL>s? It seems to me that the comm. driver
returns <NUL> to mean "nothing there, Charlie."
-Russ
#: 14698 S1/Help/Data Lib Tools
10-Jun-88 07:11:20
Sb: #14675-#xmodem - Runterm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Sysop/Betty Knight 76703,4037 (X)
Betty,
I thought I explained that. Prior to early January, our XMODEM would resend
a packet if while waiting for an ACK it either timed out or received an
unanticipated character. This was fouling up some commercial users whose
XMODEM support barfed on a duplicate packet (actually, counted it as an error
and bombed after so many of them). I changed it so that a time out and
spurious characters both resulted in an error count but did not resend the
packet. RunTerm, for some reason, begins its download dialogue with
<NAK><CR>OK, SEND THE FILE<CR>, and the stuff after the <NAK> was counted as
errors. After ten errors, ("<CR>OK, SEND "), we aborted and sent three <CAN>s
(control-X). Level 5 is supposed to fix this by counting timeout as an error,
but spurious characters are ignored. I suspect that the level 5 version of
FILTRN on your host is not current. I'll make sure the correct version is
propogated everywhere.
I just checked. I don't know what version you were running, but Level 1 now
has (114) on your host. Give it a try.
-Russ
#: 14709 S1/Help/Data Lib Tools
10-Jun-88 18:54:44
Sb: #14697-XMODEM - RunTerm
Fm: Ed Flinn 73127,1476
To: Russ Ranshaw (CIS) 70000,1010
Russ,
Nope, these were type:/bin downloads. I did try an /asc tonite, just to
see what it would look like, and the $00FF string gets translated to, I think,
$707F401F. Neither the 00 nor the FF get thru clean.
I don't know enough to discuss how term pgms for the Commodore are able to
differentiate between a NUL and nothing.
I started doing another test tonight of the download problems, and, having
seen your message to Betty, started querying the VER. I believe that at one
point, I had version 1B(114), and that in this version, both LEV 0 and LEV 5
were dropping NULs. I cannot, however, duplicate this, and so mention it only
in passing.
Once I began a more formal recording of my tests and results, the only VER
I was getting was 1B(101). In this version, three term pgms that I consider
solid worked perfectly at level 0 and level 5, as did the patched Softex
Runterm. The unpatched Softex Runterm aborted the download at both levels.
<ed>
#: 14713 S1/Help/Data Lib Tools
10-Jun-88 23:15:27
Sb: #14697-XMODEM - RunTerm
Fm: Todd Heimarck/Sysop 76703,3051
To: Russ Ranshaw (CIS) 70000,1010
Russ:
A null string (nothing there) is not the same as a <NUL> character, which
has an ASCII value of 0 and a string length of 1. Of course, you already knew
that. In Commodore 64 BASIC, if A$ is a null string, LEN(A$) will be zero and
ASC(A$) doesn't work. On the C-128, LEN is zero and ASC(A$) is also zero.
Also, if you're inputting from a disk or from the RS-232/modem channel, if
nothing's there, you'll get a null string, which could be interpreted as a
<NUL> character.
However, there are ways (status variables) to check to see if you really got
a character. After I wrote that other message to you, I found out that RUNterm
was written by someone whose programming abilities are excellent. He wouldn't
make the mistake of interpreting a null string as a <NUL> or vice versa,
although he did make a dumb mistake when he decided to send that greeting. He
probably tested it on CIS and it worked on the old software.
ToddH
#: 14714 S1/Help/Data Lib Tools
11-Jun-88 07:34:07
Sb: #14696-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010
Dear Russ,
Thanks for your messages. I'm sorry I didn't think to check the Version
of FILTRN on LEV 5. Friday I went to my PERsonal Area and it so happened that
sometimes in FILTRN I got 1B(114) and sometimes it would be back to the LEV 0
version 1B(101). I tried to find a pattern of why this was happening but
that's something else.
I think Ed Flinn is correct about the LEVel 5 1B(114) being a problem. I
went to level 5 and FILTRN 1B(114) using my regular "VT100-128" Terminal
Program and downloaded my test Binary file with its XMODEM. It downloaded as
seven XMODEM blocks but it could NOT be "loaded and run." Then I followed the
same procedure using my C64 VIDTEX4.2 and the file downloaded and "loaded and
ran" just fine. That's why I say there is a problem with FILTRN 1B(114) and
XMODEM (Commodore). Also today, Saturday 06/11/88, I did another test using
the "Common Sense" Terminal program and XMODEM to download my test Binary file.
It downloaded but would NOT "load and run" on 1B(114) but the same test on
1B(101) downloaded and ran just fine. I'm really concerned if the FILTRN
version 1B(114) becomes the ACCESS software because no Commodore Members could
download with XMODEM and that would be a disaster. Now back to the "RunTerm"
XMODEM problem.
<continued on next message>
#: 14715 S1/Help/Data Lib Tools
11-Jun-88 07:37:14
Sb: #14696-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010
Dear Russ,
<continued from previous message>
First, I feel certain that my testing on Thursday was with 1B(101) and NOT
1B(114). I tried your procedure of using a CTRL-U and it did display one
block (128 characters) and a CTRL-X started the "RunTerm" XMODEM download. This
download could NOT be "loaded and run." I found also that on 1B(114) I did NOT
need the CTRL-U -- it would download with no timeout. However, again it would
NOT "load and run." This was with the older version of "RunTerm." The SOFTEX
version of "RunTerm" did need the CTRL-U to get it started but after
downloading it also would NOT "load and run."
Ed tried several Commodore XMODEM Terminal Programs and none of them
downloaded correctly as PRG files with 1B(114). I have one other Terminal
Program (Bobs Term 128) that I will test with both 1B(114) and 1B(101). I
think I know the outcome will be like all of the others -- I'll let you know.
Does the FILTRN software become the ACCESS software? I'll do some further
testing and will look forward to any new information you can supply. Thanks and
take care.
Sincerely,
Betty Knight
#: 14716 S1/Help/Data Lib Tools
11-Jun-88 07:48:26
Sb: #14709-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Ed Flinn 73127,1476 (X)
Dear Ed,
I think you are correct in that besides the "RunTerm" problem
there are XMODEM problems on 1B(114) with quite a few CBM XMODEM
Terminal Programs.
See my messages #14714 and #14715 on my testing. Thanks for your
research on this problem. Take care.
Sincerely,
Sysop/Betty Knight
#: 14721 S1/Help/Data Lib Tools
11-Jun-88 17:32:10
Sb: #XMODEM - RunTerm
Fm: Bob Raikes 75616,1153
To: Betty Knight 76703,4037 (X)
I've been scrolling hang-jawed through the messages about the RunTerm
XModem problem. It's too bad this forum has been cursed with this
thing, but it may help some users. Which brings me to the point: Is
there any way of communicating with RunTerm users of other forums --
and users of other software that may have been caught in the same jam,
if any? It'd be nice to spread Thomas Fry's patch as far as possible.
Also, I wonder if anyone has mentioned this to the source of the
program. Thanks for working on this. Bob R.
#: 14723 S1/Help/Data Lib Tools
11-Jun-88 19:03:46
Sb: #14526-XMODEM - RunTerm
Fm: John F Davis 73455,43
To: Todd Heimarck/Sysop 76703,3051
Re: 8th bit, X-modem and Mr. Farmer's comments.
I belive that what I said is X-MODEM requires 8 bit word length.
Most terminal programs can use either 7 or 8 bit but X-MODEM needs 8
What I should have made clear is 7-8 bit word length is for TEXT only.
Not for data transfer. Oh well. That causes more than one problem
beacuse Comp-U-Serve will work with either 7 or 8 bit word length
except for file transfer
Well, Hope this clears things up
John
#: 14724 S1/Help/Data Lib Tools
11-Jun-88 19:05:59
Sb: #14721-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Bob Raikes 75616,1153
Dear Bob,
We have quite a thread going in CBMART on XMODEM in General on a
new version and also with two versions of RunTerm. Thomas Fry gave
the patch for what we are calling thw old version and Sysop Todd
Heimarck gave the patch for the newer SOFTEX version. And they both
worked fine. Sysop Todd mentioned we need a general help file on this
for our users, so in CBMART's LIB 1 there is a file named RUNTER.HLP
which explains the patches and how you find out if you need them and
which "poke" patch to use. It is a very short file and please, if you
want to recommend interested Members to this CBMART LIB 1 file we would
be glad to help them,.
Also, all of the Messages on this XMODEM situation has been
captured in a file and is available for all members to read or download
(it's rather long so downloading would be better for printing off
line). While this is a continuing message topic, every day I put the
new messages in the XMODEM.THD file in CBMART's LIB 1.
I'm having fun working on testing with new versions at other
LEVels so we can be certain the new ACCESS Software, when it gets here
will be working with Commodore's XMODEM. The Sysops and Ed Flinn have
been working with Russ Ranshaw for information. If you have anything
to add to this thread, please leave any questions about it and heep the
"thread" going to completion.
So, Bob, you could recommend your friends read the two files in
CBMART's LIBrary 1 named:--
RUNTER.HLP
XMODEM.THD
Thanks for your help and take care.
Sincerely,
Sysop/Betty Knight
#: 14725 S1/Help/Data Lib Tools
11-Jun-88 19:16:56
Sb: #14721-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Bob Raikes 75616,1153
Dear Bob,
I'm delighted we can help our Members with the "RunTerm" problem
and that we can accumulate this nice "thread" of messages (XMODEM.THD)
concerning another possible XMODEM problem.
Tell your friends about the files and then they will know where to
get some answers. Good luck and take care.
Sincerely,
Sysop/Betty Knight
#: 14740 S1/Help/Data Lib Tools
12-Jun-88 20:18:24
Sb: #14675-#xmodem - Runterm
Fm: Neil Cumfer 72457,111
To: Sysop/Betty Knight 76703,4037 (X)
Dear Betty,
I have been following the messages about Xmodem very closely because
I too have experienced problems downloading from CompuServe to my B128
(though not recently since my current terminal program has a better
implementation of Xmodem).
I agree with you, CompuServe's Xmodem will not work properly under
Level 5 (version 114). I uploaded a short BASIC program to my personal
area. Don't know if you are a programmer, but Commodore BASIC uses NUL
to separate each program line. NUL's are also possible for the bytes
reserved for line numbers and line links, and my program had several of
them. When I downloaded it with version 101 after issuing the R FILTRN
command, it came across perfectly. But when I changed to level 5 and
used version 114, it was another story. I did get a downloaded file on
my disk, but as another member wrote, it did not have any NUL's in it
so of course it would not run. All the other 7-bit and 8-bit characters
(including BASIC keyword tokens) were identical to the original file.
My file, after the Xmodem padding was added in, had 768 8-bit bytes.
Curiously, version 101 reported the size as 384 characters, while
version 114 said it had 768 characters. Also, version 101 transmitted
the padding as NUL's, which is what my terminal used when uploading the
file. But version 114, after deleting the NUL's, was 37 (8-bit)
characters short of an Xmodem block and so it padded the file with
^Z's.
I am surprised that CompuServe did not catch this error until you
reported it. Surely Commodore isn't the only computer that uses the
NUL?
Good luck in getting them straightened out! I wonder if FILTRN
means FILTeR out the Nulls?
Neil Cumfer
#: 14743 S1/Help/Data Lib Tools
13-Jun-88 06:38:21
Sb: #14696-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010
Dear Russ,
I completed another test in FILTRN both versions. I used "Bobs
Term Pro 128" and its XMODEM protocol. This is a Terminal Program
which many use and never have any trouble downloading with XMODEM.
Using my test Binary file, I downloaded it from LEVel 5 Version
101 and the download took place; then the program "loaded and ran" just
fine.
Using the same procedure I downloaded from LEV 5 Version 114 and
the file downloaded, BUT it would not "load and run". So this is yet
another Commodore Terminal Program which will NOT download with XMODEM
from 1B(114). Like I mentioned before - if FILTRN 1B(114) is ever
moved to the Forum ACCESS software we will not be able to download from
our LIBraries with XMODEM with Commodore personal computers.
I won't test anymore until I hear back from you on 1B(114).
Thanks and take care.
Sincerely,
Betty Knight
#: 14743 S1/Help/Data Lib Tools
13-Jun-88 06:38:21
Sb: #14696-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010 (X)
Dear Russ,
I completed another test in FILTRN both versions. I used "Bobs Term Pro
128" and its XMODEM protocol. This is a Terminal Program which many use and
never have any trouble downloading with XMODEM.
Using my test Binary file, I downloaded it from LEVel 5 Version 101 and
the download took place; then the program "loaded and ran" just fine.
Using the same procedure I downloaded from LEV 5 Version 114 and the file
downloaded, BUT it would not "load and run". So this is yet another Commodore
Terminal Program which will NOT download with XMODEM from 1B(114). Like I
mentioned before - if FILTRN 1B(114) is ever moved to the Forum ACCESS software
we will not be able to download from our LIBraries with XMODEM with Commodore
personal computers.
I won't test anymore until I hear back from you on 1B(114). Thanks and
take care.
Sincerely,
Betty Knight
#: 14744 S1/Help/Data Lib Tools
13-Jun-88 08:52:34
Sb: #14709-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Ed Flinn 73127,1476 (X)
Ed,
Thanks. I still don't know why RunTerm is aborting, but I do know whay
<NUL>s (and <^Z)s are being dropped. The correction has been made for this and
should be up by tomorrow (FILTRN version 1B(115)). I assume you meant
"/type:bin" and not "type:/bin" because the latter is obscure to me. Yes,
selecting /type:asc on a binary file will yield strange results which bear
little relationship to the actual data.
Your failure to return to 1B(114) was likely due to trying the level change
from within FILTRAN rather than outside of it. It is best to return to the
"OK" prompt, issue the "LEVEL" command, then "R FILTRN". Do a VER immediately.
If RunTerm fails with 1B(114) or later, there must be something else going on
and we have to find out what.
-Russ
#: 14745 S1/Help/Data Lib Tools
13-Jun-88 08:52:39
Sb: #14713-XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Todd Heimarck/Sysop 76703,3051 (X)
Todd,
I found a glitch in the latest code on our end which drops <NULs> (and ^Zs).
It is fixed and on its way.
-Russ
#: 14746 S1/Help/Data Lib Tools
13-Jun-88 08:52:45
Sb: #14714-#XMODEM - RunTerm
Fm: Russ Ranshaw (CIS) 70000,1010
To: Sysop/Betty Knight 76703,4037 (X)
Betty,
I found and fixed the missing <NUL> problem. It should arrive in FILTRN
version 1B(115) tomorrow.
I'm still concerned, tho, about unpatched RunTerm aborting downloads.
-Russ
#: 14748 S1/Help/Data Lib Tools
13-Jun-88 14:26:41
Sb: #14746-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010 (X)
Dear Russ,
That's good news on FILTRN 1B(115). As soon as it is available I will run
my XMODEM tests again and let you know. Yes, I too am concerned about the
"RunTerm" aborting downloads. I'll test it again on 1B(115).
I'm packing to move so all of my computer manuals and most software is
packed (I cannot disassemble now). But after testing with 1B(115) perhaps
either one of the CBMART Sysops or myself can disassemble "RunTerm" (both
versions - they are written in assembler code) to check the actual code that is
causing the problem.
Thanks for the information and take care.
Sincerely,
Betty Knight
#: 14749 S1/Help/Data Lib Tools
13-Jun-88 17:49:52
Sb: #14740-#xmodem - Runterm
Fm: Sysop/Betty Knight 76703,4037
To: Neil Cumfer 72457,111 (X)
Dear Neil,
Thank you very much for your research on XMODEM; it was appreciated. When
you continue with the new messages in this thread, you will see that on LEV 5
the FILTRN version 1B(115) will be available tomorrow morning which corrects
the NUL problem. FILTRN is really FILe TRaNfer (I think) but your definition
of the acronym was very apropos to this problem <smile>.
Thanks again for your message and take care. It's nice to see you here.
Sincerely,
Sysop/Betty Knight
#: 14755 S1/Help/Data Lib Tools
14-Jun-88 08:45:46
Sb: #14746-XMODEM - RunTerm
Fm: Sysop/Betty Knight 76703,4037
To: Russ Ranshaw (CIS) 70000,1010
Dear Russ,
Congratulations - it works! Remember what I said!
It seems to me that all is well with FILTRN version 1B(115). I did my
tests with it and here are the results:
"VT100-128" XMODEM OK
"Common Sense" XMODEM OK
"BobsTerm Pro 128" XMODEM OK
"VIDTEX4.2" B-Proto OK
"RunTerm" (58 Block) XMODEM OK
"RunTerm" (46 Block) XMODEM OK
Both "RunTerm" versions were run withOUT the patch and all is well. Do
you have any idea when FILTRN 1B(115) will become the Forum ACCESS software?
The patches for "RunTerm" will be needed until the ACCESS software is
corrected. I doubt that many Commodore users use FILTRN in their PERsonal Area
for downloading files - they download from the Forum LIBraries.
Thanks for all of your help and "trouble shooting" - I enjoyed it because
it was what I used to do for a living <smile>.
Take care.
Sincerely,
Betty Knight
#: 14870 S1/Help/Data Lib Tools
29-Jun-88 11:33:42
Sb: Bon xferique...
Fm: Russ Ranshaw (CIS) 70000,1010
To: All
It seems that we have fixed the XMODEM protocol problem which was plaguing
users of RunTerm. The new version of ACCESS which contains the fix is on its
way down thru the various levels of testing and should arrive here in the not
too distant future. I will now fade away from CBMART. If anyone has any
additional problems with XMODEM or B protocols, don't hesitate to send me an
EasyPlex.
-Russ
("Who _WAS_ that masked man?")