I look into how one could escape an Atom feed using Ruby and making links absolute. The intention was to use it for fixing links in my Atom feed, but I didn't need to turn it into a plugin since the actual problem was with the main url in the Atom feed (it was missing https:// in the start).
Published: 2020-05-04
Last updated: 2020-05-31 12:27
So I had some trouble getting correct links in my Atom feed. The links are written as relative in my posts, while feed apps expect them to be absolute.
I was looking for any plugins that would transform them easily to absolute links, but fell short. I therefore started looking at how hard it would be to develop my own Ruby plugin to Jekyll. Having no previous experience with Ruby, than typing bundle install and bundle exec jekyll serve, I felt that I had some reading to do.
The following script opens up the final Atom feed XML file, prepends each url with the values base_url and page_url, and then HTML escapes the contents of the entries.
newer
require 'nokogiri'
require 'cgi'
# Parse document
doc = Nokogiri::HTML(open('atom.xml'))
contents = doc.xpath('//feed//entry//content')
# Fix URLs for images and links
contents.xpath('//img').foreach { |img|
img['src'] = base_url + page_url + img['src']
}
contents.xpath('//a').foreach { |a|
a['href'] = base_url + page_url + a['href']
}
# Escape the content of each each entry
contents.children = CGI.escapeHTML(contents.children.to_s)
The actual problem with the feed seemed to be with the atom feed URL not being absolute, which was fixed by adding the absolute_url filter to the url:
<link href="{{ page.url | absolute_url }}" rel="self" type="application/atom+xml" />
which then turns into
<link href="https://blog.cedernaes.com/atom.xml" rel="self" type="application/atom+xml"/>