Run make and then use the resulting executable revp.exe.
Note: revp also builds on Windows. Just create a new Visual Studio project and import the files.
Run server with ./revp.exe -s -l 5000 -p 6000, where 5000 is where your client will connect and 6000 is where proxy connections will be serviced.
Run client with ./revp.exe -c -a 192.168.1.1 -z 5000, where 192.168.1.1 is your server address and 5000 is your chosen server port that listens for a client.
Open up a browser in your server instance and configure your HTTP proxy to port 6000, now you will be effectively browsing the Internet with client's connection.
There are 3 components to revp: server, client, and proxy connections.
Server listens for client and starts listening for proxy connections once client connects.
All proxy connection requests are routed to the client through the server.
Proxy connections are expected to make a first request of HTTP CONNECT and once connected they can exchange any data of any protocol with the destination address:port requested in the HTTP CONNECT. Server transparently routes all traffic from and to client and client routes all traffic from and to the remote destination.
What makes this a "reverse" proxy is the fact that the client initiates connection to the server, so there's no need to open ports or write firewall rules for the client whose network access you want to reach into -- you can basically run this on any network node whose firewall allows outbound traffic to your server port (essentially all firewall configurations found in the wild allow all outbound traffic).
-h [ --help ] Help screen
-s [ --server ] Starts revp in server mode
-c [ --client ] Starts revp in client mode
-l [ --clientPort ] arg [Server mode] Port that listens for client
-p [ --proxyPort ] arg [Server mode] Port that listens for proxy connections
-a [ --serverAddress ] arg [Client mode] Address of server
-z [ --serverPort ] arg [Client mode] Port of server