On 2019-05-03 22:37, Paul Koning wrote:
On May 3, 2019, at 3:52 PM, Johnny Billquist <b...@softjar.se> wrote:
That came across wrong...
Flow control would most certainly help, however, when running in a simulator,
with a terminal window, connected to the system, there is actually no flow
control available, so you will loose data.
Why would there be no flow control available?
Flow control is part of the protocol, no different from escape sequence parsing. If a terminal
emulation receives a control/S from the host, it should stop sending characters to the host. For
that to be meaningful, the "off" action has to happen close enough to where the
characters are picked up by the driver that you don't get the simulator equivalent of "Fifo
overrun" but that is not hard to arrange.
If the emulation currently doesn't do this, forcing the character rate down to
"really slow" should be an adequate workaround.
Essentially, terminal emulators usually totally ignore XON/XOFF if it
comes on a telnet session, or pty. Unix do handle XON/XOFF if you come
in on a physical serial port, but basically not in any other
circumstances. Terminal emulators themselves don't really try to handle
this. Both ptys, as well as telnet (or any other tcp connection) have
their own flow control anyhow, so there is no need, and possibly no
point in trying to handle XON/XOFF.
This was a fun exercise I had to sort when I did telnet in RSX, as the
normal serial port flow control isn't really helpful anyway at that
point. Delays are just too long, not to mention that no telnet client
would ever understand or care if I sent it an XOFF.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: b...@softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
_______________________________________________
Simh mailing list
Simh@trailing-edge.com
http://mailman.trailing-edge.com/mailman/listinfo/simh
