That's the option I'd go for.

Barring that, as silly as it sounds -- gopherd can read mbox format. You just point the gopher server at the file, and it'll generate an index for it, and pull out individual messages on request.

I always thought httpd was a bit of a step backwards in that regard.