What I'm trying to do here is set up a minimalistic blogging platform, and learn a bit of practically-oriented Prolog in the process. I'm leaning pretty heavily on Annie Ogborn's tutorial at this point.
I really dislike doing any composition in the browser, and so what I'm after here is just a system that will take a directory of markdown files, and a table of contents file in a nice, simple syntax of its own, and take care of beautifying and serving the documents from there. Posting to the blog will just be done through SSH. I don't see any need to complicate that any further.
Maybe I'll figure out a neat way of integrating comments, etc., in the future.
In order for anything to actually work, however, you need to first provide the blog with some content. The fastest way to get started on this is just to
cp -rv content.example/ content/
And then make whatever changes you see fit.
To serve on port 8000, make sure that that port and the appropriate bind
address are specified in about.prolog, in your blog directory. Then
either link, copy, or mount your blog directory under ./content, and
run swipl -l plog.pl -g start.
Blog posts go in content/posts, and should be
in markdown
format. Information pages that aren't exactly posts, like the GPL
license document, or the "about this blog" page, go in content/info.
Each markdown file should contain a YAML header. Take a look at the posts in
content.example/posts to see the format.
But basically it should look like this:
---
title: "How I Taught Barnacles to Dance"
tags:
- barnacles
- choreography
date: 2023-01-26
abstract: "They said it couldn't be done, but every last one of them was a coward."
author: me
---
If you attach the special tag draft to a post, then it won't show up in the table of contents, but anyone who knows the URL will still be able to navigate there directly. Think of it as a privacy hedge, not a brick wall.
The CSS files should go in content/css. Use the examples given as a template,
and work from there, until things look the way you'd like.
The favicon.ico can be found in content/img, which is where any images linked
to from your posts should also be stored. The links should be relative to the
content directory.
Metadata for your blog should go in content/about.prolog. The syntax used
there should be fairly self-explanatory.
If you would like to proxy to the P'log server through Nginx, you can
add the following lines to your website's entry in /etc/nginx/sites-available,
making whatever adjustments you see fit:
location / {
try_files $uri @upstream;
}
# proxy through to the p'log server
location @upstream {
# Disables HEAD requests to the site.
# SWI-Prolog currently does not support
# head requests very well.
if ($request_method = HEAD) {
return 405;
}
proxy_pass http://127.0.0.1:8000;
# or whatever port you're using
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header Host $http_host;
proxy_set_header X-Nginx-Proxy true;
proxy_redirect off;
}
It's expected that you will be running P'log from the directory into which
you have cloned its git repository, and that the data that makes up the
meat of your blog will live in its own git repo, under ./content/. This
is shouldn't be strictly necessary at this point, but if git information is
available, P'log will use it to update metadata on the posts.
If you want to run P'Log in a Docker container, you can do so easily. Just run
the plogserve.sh script, like so:
$ ./plogserve.sh $PATH_TO_YOUR_BLOG $BIND_ADDRESS $PORT
This list is complete. You can help by adding to it.