[Simh] GOLD Keys and such

[Christian Gauger-Cosgrove]

On 21 December 2015 at 02:14,  <lists at openmailbox.org> wrote:
> Even for the mid 1970s MVS 3.8 OS IBM placed in the public domain nobody
> uses anything but 3278s except maybe for console logging (not as a general
> console). The printing terminals were useful for OS/360 console loggers
> but few Hercules users run OS/360. Bottom line and for everything else when
> you're using a terminal with MVS it's a 3278 and needs a tn3270 client if
> you are not direct attached. tn3270 covers 100% of normal use cases for MVS
> terminals except if you insist on using an optional printing console. You
> certainly don't need one.
>
The printing terminals (1052-7 or 3215; your choice) are required if
you do a sysgen from the starter system. Since the starter system is
*not* built with support for 3270 type consoles. Also, while I haven't
used it, I'm pretty sure the "Hercules Automatic Operator" scripting
necessitates using a printing console in the "-C" mode (so the script
can see the output and then respond automatically). Neither the 3215
or 1052 devices are used for TSO, however the 2703+TELE2 emulation may
be used for TSO.

So yes, printing consoles certainly aren't necessary to operation of a
fully set up MVS 3.8J system, however for the MVS 3.7 "starter" system
needed to actually sysgen 3.8 they are required (and they are required
for HAO if I recall correctly).


> Are you saying Putty's telnet doesn't support the printing terminals
> Hercules supports? Because if that is not what you are saying then we seem
> to be going around in circles. Telnet clearly won't work for a normal IBM
> terminal.
>
There is a bug in telnet negotiations in Hercules, that causes PuTTY
to be unable to connect to Hercules.


On 21 December 2015 at 02:31,  <lists at openmailbox.org> wrote:
> We should clarify the way you connect to Hercules _emulating_ those devices
> is telnet. Telnet does not work with real IBM devices on real IBM mainframe
> hardware.
>
Yes, correct. Hercules provides two methods of connecting to it
depending on the kind of device being emulated. For 327x devices you
*must* use a tn3270 client (like x3270). For 1052-7, 3215, or 2703
with TELE2 devices you *must* use telnet.

In the real world of IBM hardware you cannot telnet into the
mainframe; nor could one normally connect to the mainframe via a
"normal" terminal like a VT-100, or the like. There are some rare
exceptions, but those don't count.



So, just to recap this whole thing:
The Hercules (IBM System/370 and up) emulator provides emulation of
the following types of console/terminal which are reachable by the
following protocols:
* IBM 327x --- Local, non-SNA, 3270 device; connects via tn3270
* IBM 1052-7 --- Console printer-keyboard device; connects via telnet
* IBM 3215 --- Console printer-keyboard device; connects via telnet
* IBM 2703 with TELE2 control --- Dial-in TWX (e.g. Teletype Model 33,
or Teletype Model 35) remote line; connects via telnet

PuTTY's telnet client cannot connect to Hercules, as there is a bug in
telnet negotiations in the Hercules console.c code that makes it
non-RFC compliant.

One can get around the need for a telnet connection to either 1052 or
3215 devices by specifying a 1052-C or 3215-C; which still emulate the
device mentioned but redirect the console output to the Hercules
command line (and take input from said command line as well).


Anyway this has gone way past the point of being related to SIMH; and
in reality is a non-issue. A quick "TL;DR" as the internet neologism
would be: Hercules has a telnet negotiation bug that means PuTTY can't
connect to it.

Cheers,
Christian
-- 
Christian M. Gauger-Cosgrove
STCKON08DS0
Contact information available upon request.