Add another Gopher recipe vault.

Loop through sub-directories and files, and handle (ASCII text) files properly.

Parse out the structure of the Gopher file.

Wait for potential secondary/tertiary/etc. replies.  This happens if the reply is too big to fit in a single TCP frame.

Also restructure the code some to improve readability.  The core loops are still very unintuitive, but the rest is decent now.

First version.  Fetches a Gopher page, at least so long as there's only one server reply.