On Mar 28, 2014, at 4:28 PM, Emmanuel Lécharny wrote:

Le 3/28/14 8:56 PM, Archer, Garry a écrit :
Hello All,

In the TCP/IP client/server implementations, servers create NioSocketAcceptors 
to to listen for incoming connections and clients create NioSocketConnector to 
connect with the server.  I have used this architecture using Mina several 
times.

Now, I want to simulate a server on a machine that only has serial 
communication connectivity (RS-232).  The client is TCP/IP.  There is a 
terminal port server in between that converts the serial signals into TCP 
signals and vice-versa.

I can see how to create a Mina client on the machine simulator with 
SerialConnector, but I cannot see how to create a Mina server on the machine 
simulator.  I.e., is there something like a "SerialAcceptor"?  I don't see it 
in the API, so is there another way to create a server on a machine (or PC) 
that only has serial communications, no TCP/IP?

We haven't implemented such a class. That would be interesting to have
one...


Yes, I understand.  I don't need one for a production environment, but I wanted 
to simulate this machine that sends it's data via serial communication.  I 
suppose it's not wanted by many developers, but may have been interesting to 
have one, as you said.

Basically, I was going to develop a Mina-enabled test program listening for 
requests on the RS-232 COM port using a "SerialAcceptor."

I don't really need this "SerialAcceptor" any other time, particularly in a 
production environment.  Except this one machine, all our other machines use 
TCP/IP connectivity.

It was easy to write a Mina-enabled program to send data through the COM port 
using the SerialConnector class.

Oh well, c'est la vie.  I am currently debugging the WriteRequestQueue problem 
I had in another thread with test programs using TCP/IP connectivity in a 
testing environment.   The problem does not occur with another machine I 
communicate with using Mina in a production environment.  I tried to replicate 
that system and I do not yet know why one works with no problems and the other 
one does!

I'll find out what it is sooner or later!  I suppose that is why they pay me 
the "big bucks!" ;-)

Regards,

Garry Archer
Systems Programmer
Department of Pathology
Yale School of Medicine

Reply via email to