| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 5222.1 | | LASSIE::GEMIGNANI | | Tue Feb 18 1997 17:20 | 9 |
| Actually, it is.
The outgoing TELNET connection that you created is attempting to
initialize the connection with the TELNET protocol. It is basically
advertising to the remote TELNET server that it wants to be in
character-at-a-time mode.
The newer TELNET servers have disabled this feature, though.
Please be sure that you have the latest TNDRIVER.
|
| 5222.2 | | CSC32::J_MORTON | O8-OO-2b || ! 2b | Wed Feb 19 1997 14:05 | 6 |
| The customer sees this with V4.1 ECO 4. Is the behavior
changed later than this?
Thanks!
Jim
|
| 5222.3 | Is a later ECO needed? | CSC32::J_MORTON | O8-OO-2b || ! 2b | Thu Feb 27 1997 11:33 | 1 |
|
|
| 5222.4 | | UCXAXP::GEMIGNANI | | Thu Feb 27 1997 14:39 | 7 |
| The code should have been in SSB 4.1.
I added an undocumented /OPTIONS=(ECHO,SGA) to turn on the negotation.
Most people didn't want it to negotiate these options, just a raw data
stream.
You are definitely seeing THIS negotiation, right?
|
| 5222.5 | | CSC32::J_MORTON | O8-OO-2b || ! 2b | Fri Feb 28 1997 11:20 | 11 |
| Hi John,
What I put in .0 is what the log files show. I'm not exactly sure
what those leading characters are, just that they exist.
The customer is at V4.1 ECO 4. Is there anything else we can
do to verify what these characters are? or eliminate them?
Thanks,
Jim
|
| 5222.6 | | UCXAXP::GEMIGNANI | | Fri Mar 07 1997 19:06 | 26 |
| It turns out that all of these things are based upon the TELNET
protocol. This means several things:
1) Special command sequences are transmitted as <255><...>.
The TELNET RFC covers these.
2) REAL <255>s are sent as <255><255>, that is, they are doubled.
3) Lines terminated as \n (newline character, LF) become CR/LF
pairs and real CRs sent alone are <CR><NULL>.
FTP, SMTP, and others are based upon the TELNET protocol, while they
don't provide a TELNET service.
Is the application sending <255>X (which is being expanded to
<255><255>X per the description above)?
I am assuming that the X is the character sent by the user's program.
Keep in mind that the TELNET protocol sends <IAC><option> [IAC is 255,
and means Interpret-As-Command].
Be sure that the customer knows that the TNAs speak TELNET protocol.
A future release of UCX will give you the option to create TELNET or
RAW TN devices (with or without TELNET protocol).
|
| 5222.7 | more info. | CSC32::J_MORTON | O8-OO-2b || ! 2b | Mon Mar 10 1997 12:55 | 49 |
| Hi John,
I'm not sure what that "X" is either. The customer sent the log along with
the command procedure in .0.
Below is a procedure I'm using and my SMTP log. I do the following:
$ DEFINE/SYSTEM UCX$SMTP_PROTO_TRACE 1
$ DEFINE/SYSTEM UCX$SMTP_PROTO_DEBUG 1
$ UCX STOP MAIL
$ UCX START MAIL
$ @<the_procedure>
fyi/thanks,
Jim
-----------------
$ set noon
$ read sys$command tn/prompt="Enter TN #: "
$ lf[0,8] = 10
$ cr[0,8] = 13
$ telnet/create stitr 25 'tn
$ set term tna'tn: -
/notype_ahead/unknown/hostsync/ttsync/nointeractive/nowrap/eightbit/hard-
/pasthru/nobroadcast/notab/noecho/nodisconnect/nofull/nohangup/perm
$ open/append rtest TNA'tn:
$ write rtest "QUIT"
$
$READ_LOOP:
$ read/time=2/error=READ_DONE rtest output
$ write sys$output output
$ goto READ_LOOP
$
$READ_DONE:
$ wait 00:00:5
$DELETE_PORT:
$ if f$trnlnm("RTEST") .nes. "" then close rtest
$ telnet/delete 'tn
$ exit
-----------------
. . .
#smtp_read_reply:recv:status=26, vaxc$errno=1
recv buf=\ffffffff\fffffffd\18\d\aQUIT\dData:\9\d\a\0\dData:\9
#smtp_read_reply:B4 ucx$select: vaxc$errno=1, errno=65535
. . .
|
| 5222.8 | | LASSIE::GEMIGNANI | | Tue Mar 11 1997 13:18 | 17 |
|
Yeah, I found it.
That's IAC/DO/TTYPE -- the TNA wants the remote system
to do terminal type negotiation. The driver does this
as part of the initialization of the created session.
Keep in mind that this is a TELNET connection speaking the
protocol, and this is valid TELNET protocol. One solution
is for the customer to treat this as protocol and either
NOT respond -or- respond with IAC/WONT/TTYPE (<FF><FC><18>).
The other solution is to submit a priority 3 IPMT (making it
a nicety, not a necessity). I have done similar stuff for
SGA and ECHO (which, you may have noticed, doesn't come out
unless you enable it).
|