The use of control socket can create confusion as seen in #13397
control port is more appropriate in this instance https://gitweb.torproject.org/torspec.git/tree/control-spec.txt
The use of control socket can create confusion as seen in #13397
control port is more appropriate in this instance https://gitweb.torproject.org/torspec.git/tree/control-spec.txt
87 | @@ -88,7 +88,7 @@ for normal IPv4/IPv6 communication, use: 88 | 89 | ## 3. Automatically listen on Tor 90 | 91 | -Starting with Tor version 0.2.7.1 it is possible, through Tor's control socket 92 | +Starting with Tor version 0.2.7.1 it is possible, through Tor's control protocol
Please remove the trailing whitespace
2018-09-17 17:56:07 mdl(pr=14243): doc/tor.md:91: MD009 Trailing spaces
A TCP socket ("network socket") is also a socket. To be precise, the "Tor control protocol" runs over a socket, which can be of different types. On Windows this only supports network sockets, on UNIX it can use a local socket.
To be honest, I'd prefer to add UNIX socket support instead, if you think that results in confusing. I don't think changing this word is worth it.
@laanwj agree on the use of word socket however this is a different scenario It is the control protocol of tor that has an API, the control socket is just a socket through which the control protocol is used. It is akin to saying RPC API instead of Bitcoin API because you're using the API through RPC.
It is just a nit and I don't care either way but I think it should be changed for rigorousness.
I also subscribe to having unix socket support added to Bitcoin's tor control given that Debian comes with it by default and has tcp sockets disabled.