Expand description
An implementation of IPC locks, guaranteed to be released if a process dies
This module implements a locking server/client where the main cargo fix
process will start up a server and then all the client processes will
connect to it. The main purpose of this file is to ensure that each crate
(aka file entry point) is only fixed by one process at a time, currently
concurrent fixes can’t happen.
The basic design here is to use a TCP server which is pretty portable across platforms. For simplicity it just uses threads as well. Clients connect to the main server, inform the server what its name is, and then wait for the server to give it the lock (aka write a byte).