Redis recently introduced support to talk the Gopher protocol. Yes, that protocol that in the '90s used to be the alternative to the web and now is almost forgotten. You can read more info about Redis and Gopher here in the antirez's blog post.

For the documentation about the Redis Gopher implementation please check the Redis official doc, specifically the Gopher protocol page.

However the raw Redis implementation is able to serve Redis keys using the Gopher protocol. If you want to create your own "Gopher hole", i.e. your site with your own content served via the Redis Gopher implementation (in that specific case at least), you need some kind of authoring tool.

This Ruby script, gopher2redis.rb, converts a directory and its subdirectories, with the text, gif, and binary files it contains, into a Redis key structure that can be served via the Redis Gopher implementation.

Gopher sites are listings that point to other nested listings of resources or to files that can be loaded as text-only documents to read, or as binary files to download or GIF images to visualize.

To create your Gopher hole (i.e a site inside the Gopher universe), create a directory somewhere, with subdirectories if you want to have some tree-like organization of your content, which is the Gopher way.

Every file that contains a "-" character inside, like 0000-Who_Am_I.txt, will be part of the content generated in the Redis server. The part before the "-" is used to sort the listing, and is not shown in the final Gopher site. The underscores are converted into spaces.

For instance if I put the following files:

0001-My_Document_is_here.txt
0000-README.txt
0003-cat.gif

What I'll see in the final Gopher site is a listing like this:

README.txt
My Document is here.txt
cat.gif

Sometimes it is useful to show the content in reverse order, for instance you may take a blog in your Gopher hole (that is a phlog in Gopher slang), and start every post with the date like:

2019.02.05-5 Feb 2019: Programming_in_Ruby.txt
2019.02.04-4 Feb 2019: Today_strange_journey.txt

and other files with similar names. In a blog it makes sense to show the last written post as the first entry. To have such effect create an empty file called REVERSE in the directory if you want the listing to be sorted in the reverse order.

The extension of the file selects the "type" of item in the Gopher menu for that file: if it is a text file, a binary file, a gif, an image and so forth. Here is the current map:

• zip, bin, gz, tgz, o: binary files
• gif: GIF file
• jpg, jpeg, png: Image file
• html, htm: HTML file
• link: Gopher link file (see later)
• info: Some text that will be put inline into the listing as "info" fields
• i: Same as .info but the filename itself is rendered as "info" field
• All the other extensions, such as txt, c: plain text files

By default everything is a text file if not expressly specified by a different extension. In Gopher text files are the content users want to see.

Sometimes you want to link other Gopher sites from your site. In order to create a Gopher link in your listings, create a file named 0004-Also_visit_this.link site (.link extension) containing a Gopher URI like:

gopher://gopher.somesite.xyz

Or

gopher://gopher.somesite.xyz:99/0/resource

And so forth. If you specify a given resource in another Gopher server, make sure that the path includes the resource type, like the /0/ part in the URI above.

You can create all the nested directories you want: they'll be processed recursively and create other nested menu items.

Typical usage is like so:

./gopher2redis.rb --host 127.0.0.1 --port 70 \
                  --root /Users/Alice/mygopherhole \
                  --localhost gopher.alicesite.net --localport 70

The --host and --port options tell the program how to connect to Redis in order to change its content (WARNING: don't write to the wrong Redis server for an error).

The --root option is used in order to specify which directory (and nested directories) you want to load into Redis for your Gopher site.

Finally the --localhost and --localport options are used to tell the utility in which host and port the public Redis Gopher service will run, so that the utility can generate the Gopher references in the listings to point to the right hostname and port.

For instance if your public gopher server will be gopher.myserver.xyz, specify that. To test this program locally use localhost.

For other options please use the --help switch.

This repository contains an example-gopherhole directory that contains a few files that will render in a site you can display with a Gopher client. You can explore the directory to check how it is assembled. Follow these steps if you want to render the site into your local Redis instance and view it with lynx as the Gopher client:

• Start a Redis instance, latest unstable branch (or any Redis version 6 if already released -- at the time of writing it's alpha code), with the --gopher-enabled yes option in the command line, or with the same option inside redis.conf.
• Translate the example Gopher hole directory into the Redis dataset using the gopher2redis.rb script: ./gopher2redis.rb --host 127.0.0.1 --port 6379 --root ./example-gopherhole --localhost localhost --localport 6379
• See the result with lynx gopher://localhost:6379

Note that this time we are using port 6379 which is not the default Gopher port. If you want to use the default port, that is 70, run Redis using such port (but you have to run it as a root user, or with alternative methods to give the process access to the lower TCP ports), and change the above command line as needed. If you plan to run such setup in production, make sure to read the next section about securing your environment.

Warning: even if in Gopher mode, Redis will continue to serve normal commands: make sure to set a password using the requirepass option. Also make sure to set the gopher-enabled option to yes in the Redis configuration, otherwise the server will not talk the Gopher protocol.

Once you set a password in your Redis instance, you can use the --pass gopher2redis option in order to write the new site content using the Redis clients authentication.