It's not that the escape sequence is long, it's that the M10x acts upon
the escape sequence as soon as it is received and fails to send an XOFF
while it is responding to that escape sequence. It's almost as if the
machine has to wait until the screen is done with the command. In that
brief moment, the buffer can overload. I don't know for certain that
this is true, but when I was working on sending escape sequences to the
M10x and 200 with mComm, I found that all the overflow problems cleared
up when I added a slight delay after sending an escape sequence. On the
flip side, when sending a 15k document to the m100 through TEXT, I don't
have to have any delays since the screen is out of the loop.
Kurt
On 7/18/2015 8:34 PM, MikeS wrote:
Hmm, interesting; that would be the case in a 'normal' comm session as
well though, wouldn't it?
Assuming you're checking for an XOFF after every character sent you
should be able to handle an 'overload' of 24 characters, no? That
would be an unusually long escape sequence...
Steve spent some time a while back trying to expand the M100's receive
buffer but it turned out to be more complicated than it seemed at
first and not worth spending more time on; since you need some kind of
external hardware anyway to connect to the net it makes more sense to
handle as much of the 'protocol' as possible externally as you're doing.
I use a couple of Lantronix servers ('modems'):
http://www.lantronix.com/device-networking/external-device-servers/uds-10.html
The size of the packets sent to the RS-232 port can be set as low as 1
character and I've never had any overflow issues, but then again I
haven't sent any long escape sequences.
Definitely something to investigate, but if it works, hey...
Thanks; I'll try to find time ASAP for a bit of testing.
m
----- Original Message ----- From: "Kurt McCullum"
<[email protected]>
To: "Model 100 Discussion" <[email protected]>
Sent: Saturday, July 18, 2015 6:27 PM
Subject: Re: [M100] Telnet Success
Mike,
The overload happens when you send an escape sequence to the ModelT.
When I was writing mComm, I found that a sequence of escape sequences
for say, clear screen, move to home, reverse video ect would be
executed before sending and XOFF. So if you are sending data rapidly
during that time, it overloads the serial buffer. So a slight delay
after every escape sequence keeps the ModelT happy.
Alpha version.... let me see what I can put together. It's just a
stand alone bridge right now. I'll put some controls on the page so
you can enter a web site and pick a com port. right now that's all
hard coded.
Kurt
On 7/18/2015 3:12 PM, MikeS wrote:
Sounds very interesting!
Not sure what you mean by "overloading ... when escape sequences are
encountered"; are there issues specific to escape sequences or are
you talking about buffer overflow when XOFF is not acted on quickly
enough?
Care to send me an alpha version to play with?
m
----- Original Message ----- From: "Kurt McCullum"
<[email protected]>
To: "Model 100 Discussion" <[email protected]>
Sent: Saturday, July 18, 2015 5:40 PM
Subject: [M100] Telnet Success
I finally have some prototype code working that bridges a serial
connection to telnet and handles the overloading of the ModelT com
port when escape sequences are encountered. I've used it to log
into some BBS systems but unfortunately the 40 character screen is
an issue even on the T200. It's still a work in progress and very
rough at this point but I will keep everyone posted on the
progress. At some point I hope to add this code to mComm to give
users access to some of the Telnet systems out there.
Kurt
