simulator: clean shutdown of socket

[NickeZ]

If we don't close the socket we have to wait a couple of minutes until the OS cleans up after us. This is quite annoying because then you can't launch the simulator with the same port immediately after you kill it.

[asi345]

I tested it. When the python client first quits and then simulator is given CTRL-C, it works really well. I can instantly restart a new communication at the same port.

However, when the simulator first gets CTRL-C without quitting from python client, simulator gets stuck and can not quit until python client quits.

If this is expected and okay for you, I can approve.

[benma]

utACK

[NickeZ]

I made some attempts at a clean shutdown when a client still is connected. I think the client has to do a proper shutdown when it detects that the server is down for the OS to release the port, so we might have to do some changes to the client as well. Anyway, it seems complicated enough that I can do that in a follow up PR. This PR still improves the situation when no clients are connected.

[Noxet]

You should avoid using non-reentrant functions, like printf, in signal handlers.
The only functions that POSIX guarantee to be async-signal-safe, e.g. write(), are listed here

[NickeZ]

You should avoid using non-reentrant functions, like printf, in signal handlers. The only functions that POSIX guarantee to be async-signal-safe, e.g. write(), are listed here

addressed in #1353, thanks for the feedback!