Ordinarily, I would ask these kinds of questions of people at HP directly,
but everyone I normally talk to seems to be out of the office and on vacation.
My question is: what is the current status of the "advanced telnet" patches
for MPE/iX 6.0 and 6.5? There were named patches for these two operating
systems that corrected a number of flaws in the telnet implementation as well
as adding in the advanced telnet and XON/OFF flow control features, and a
number of you received these patches. However, shortly after they were
released and shipped, James Hoffmeister wrote this to the list about half a
year ago, on December 22, 2000:
======================================
Hello Folks @ 3000-l,
Re: DTC-less
----------------------------------------------------Ken Sletten writes--
The number of host-based TELNET sessions we run on our 959-400 is small
enough that it's not a fair test of anything (all end-users access the
3000 with our in-house client); I have never noticed TELNET putting any
significant load on our 3000.... I will just add that I *think* the 6.5
EXP/PP-2 TELNET includes support for Advanced TELNET.. For QCTerm
users, this is another good reason to go to host-based...
Ken Sletten
------------------------------------------------------------------------
FYI: The "Advanced TELNET" functionality has been pulled from all recent
Telnet patches due to a corner case of data loss. This functionality
will be returned to Telnet when the data loss is addressed.
========================================
To which Ken Sletten responded:
========================================
James said "not so fast":
> FYI: The "Advanced TELNET" functionality has been pulled
> from all recent Telnet patches due to a corner case of data
> loss. This functionality will be returned to Telnet when the
> data loss is addressed.
Shoot.... well, clearly don't want 3000 users to lose any of their
data; better later and right than sooner and wrong, I guess....
May we depend on you to give this list notice when a TELNET
fix patch that solves this "corner case" bug is available ??.. :-)
thanks for update,
Ken Sletten
=======================================
and James responded:
=======================================
Hello Kenneth,
Re: RE: Advanced TELNET pulled from patches (was "DTC-less")
I do not have a current time frame for Resolution. I understand
this is a problem Jeff Bandle turned up in Lab testing of the
Telnet/iX product.
I will update the 3000-l when we add this functionality back into
a 6.0 and 6.5 patch.
========================================
Unfortunately, I don't think the announcement that the advanced telnet
functionality was back in the 6.0/6.5 patches was ever made. Or at least, I
missed it.
BTW, there never was a problem with data loss with advanced telnet. Let me
explain the sequence of events as I understand them. The entire episode
stemmed from a misinterpretation of symptoms.
The problem came to be called "the flying B's" problem. What had occurred was
an organization that grew up on UNIX boxes was attempting to port their
software over onto the HP3000. In this software, they put the terminal into
"transmit function" mode. In this mode, a down arrow key is transmitted to
the HP3000 as an <esc>B sequence, where it can be processed.
Because of their particular combination of a speedy HP3000, a moderate-speed
LAN and a relatively slow terminal key repeat rate, the software they were
porting worked until they loaded the advanced telnet patch. Instead of now
seeing the down arrow key promote the scrolling of their software, what they
now saw on the screen were occasional "flying B's" (the last half of the
<esc>B sequence). The logical conclusion they jumped to was that the advanced
telnet patch was broken and was causing data loss. They reported this to HP,
who in turned yanked the advanced telnet patch until the mechanism of the
data loss could be determined.
What in fact occurred was that the complaining organization had forgotten (in
fact, had never) turned on typeahead in MPE. As a consequence, they were
working in the standard HP3000 input mode where every telnet character input
was absorbed into the input buffer, and once absorbed, the HP3000 goes silent
until it can process that character, before it issues a subsequent DC1,
indicating that the HP3000 is ready for the next input.
Just by chance, their HP3000 was powerful enough to actually process the data
quickly enough to actually absorb and process all of the <esc> and B
characters individually as they were being generated by the PC terminal
emulator. If they had been on an older and slower HP3000, they would have
realized immediately that they couldn't use a standard HP3000 input procedure
for this type of functionality without going to typeahead, where all of the
characters are put into a buffer and processed as quickly as they can be.
When they loaded the advanced telnet patch, which fixed many things beside
adding the advanced telnet capability, the HP3000 was now just slow enough to
be beyond the cusp of being capable of processing this stream of data as it
arrived, thus the "problem" of losing some data "appeared" because some of
that stream was hitting the HP3000 during its silent period. Putting their
code into "typeahead" mode fixed the problem, which is the manner it should
have been written from the beginning.
Nonetheless, after all of this, I now don't know the status of the 6.0 and
6.5 patches. I do know that all of the advanced telnet code is now native to
MPE/iX 7.0. But is it back in the 6.0 and 6.5 patches? Several HP3000 user
organizations have asked me for the names of the 6.0/6.5 patches for
"advanced telnet" but I am unsure of their status at the moment, and as a
consequence, I am also unsure of the correct patch ID numbers.
If anyone has that information, I would greatly appreciate it if you wouldn't
mind passing it along.
Wirt Atmar
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
