An honest Finger protocol server

Gigi is a Finger protocol server with few features–the way a Finger server should be.

• Gigi can respond to Finger requests statically.
• Gigi can respond to Finger requests dynamically.

A live, production deployment of Gigi exists at finger://fuwn.me. Feel free to poke and prod at it as you wish.

You can use a Finger client like finger to send requests, or you could use the old reliable telnet or nc commands.

finger fuwn.me
telnet fuwn.me
echo | nc fuwn.me 79

To run Gigi from a single command using Nix, execute nix run github:Fuwn/gigi; otherwise, try one of the command combinations below.

# Clone the repository locally
git clone [email protected]:Fuwn/gigi.git

# Navigate into the local repository
cd gigi

# Build and run Gigi without Nix; requires Go to be locally available
go build gigi.go && ./gigi

# Build and run Gigi with Nix
nix build && ./result/bin/gigi

This command runs the latest Gigi Docker image with port 79 mapped from inside the container to port 7979 on the host system. In practice, you'd actually map port 79 to port 79, but that requires root privileges, so we're using 7979 here.

The command also mounts the ./.gigi directory from the host system to the /gigi/.gigi directory inside the container. This is where you'd place all your profile files. In practice, you'd likely make this a named volume or mount from a more collected volume storage facility, and add files to that mounted volume path itself.

# Run Gigi using Docker with a local volume directory
docker run -v ./.gigi/:/gigi/.gigi -p 7979:79 fuwn/gigi:latest

# Run Gigi using Docker with a named volume and a remote volume directory
docker run -v gigi-data:/gigi/.gigi -p 79:79 fuwn/gigi:latest

The second command is the more practical one, as it uses a named volume to store the profile files. The named volume is persistent, and can be found at /var/lib/docker/volumes/gigi-data/_data on most semi-FHS compliant systems by default.

Docker also significantly reduces the risk of running Gigi, as it is sandboxed from the host system. In static mode, there is little to no risk, but in dynamic mode, there is a small risk for arbitrary code execution depending on your .gigi/do file.

Gigi is configured through the ./.gigi directory.

Dynamic response mode is disabled by default as dynamic code execution can be a big security risk. If you wish to live on the edge, pass the GIGI_DYNAMIC environment variable with a value greater than 1 to Gigi. Dropping Gigi into a container is significantly safer than running it on a host machine, so consider that as your primary deployment option.

Dynamic mode runs any and all executables located at the path ./.gigi/do, and passes any arguments from the Finger request to the executable.

Static mode is enabled by default. A Finger request for test will return the contents of ./.gigi/test. A Finger request of nothing will return the contents of ./.gigi/default. The default file is also the fallback file in case the requested file does not exist.

To emulate dynamic mode, minus the support for arguments, you can setup a service of some kind to periodically update the contents of one of the static files.

You can additionally modify the GIGI_PORT environment variable to change the port Gigi listens on. The default port is 79. If you're running Gigi in a Docker container, you can ignore this variable and map any ports using Docker directly.

This project is licensed with the GNU General Public License v3.0.