Possible bug in handle() method...

[eruivo]

Hi,
First of all thank you for your library, it is just what I was looking for to incorporate in one of my projects.
I have however detected that if an active connection from a client happens to go down before or without properly disconnecting the server will always drop any new connection attempt.

**if (!started || serverClient) {
  telnetServer.available().stop();**
} else {
  serverClient = telnetServer.available();
  if (strlen(welcomeMsg) > 0) {
   serverClient.write((const uint8_t*)welcomeMsg, strlen(welcomeMsg));
   serverClient.flush();
  }
}

For now I have just made a change to stop/disconnect any active connection if a new one comes in but I think implementing a simple watchdog timeout on active data transfer to stop/disconnect would be a better solution.
If the intended behaviour is as it currently is, please forgive me and ignore this.

[Wolfgang Mattis]

Hi,
this is not a bug but a problem. If there is no traffic on a telnet connection the server has problems to detect a disconnected client. It will happen but it needs a very long time. To get rid of this problem I implemented a ping function which sends periodically a null character to the telnet client if there is just no traffic. So a disconnected client will be detected very fast. This should solve your problem.
I did also a lot of other extensions/improvements (see README.md), so the external RingBuffer component is no longer neccessary. But you have to add
TelnetSpy SerialAndTelnet;
to the beginnig of your script, because it is no longer globally defined. So its now also possible to create the TelnetSpy object only if it is neccessary.

[eruivo]

Hi,
Thank you for your reply and time to look into the situation and I'm sorry for referring to it as a Bug.
I also will be checking out your latest version and the new features, I particularly like the fact that you have eliminated some of the external dependencies and the need for a global definition.