Creating a generic rss feed template, to serve up all the posts in all the Jekyll collections.

All the jekyll collections are specified in _config.yml, and all the collection-related posts and pages will be in the respective collection folders, where the collection name (aka label) is prefixed by an underscore.

Create the top-level rss file, feed.xml

layout: null
{% include all_feed.xml %}

and create the file, _includes/all_feed.xml

<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="" xml:lang="{{ site.lang }}">
    <title>{{ site.title }} by {{ }}</title>
    <link>{{ site.url }}</link>
    <copyright>All rights reserved.</copyright>
    <description>{{ site.description }}</description>
    <atom:link href="{{ page.url | absolute_url }}" rel="self" type="application/rss+xml" />
      {% assign posts = site.documents | where:"layout", "post" | sort: 'date' | reverse %}{%
      for post in posts limit:10 %}{%
        assign collection_prefix = "" %}{%
        if post.collection != "posts" %}{%
          assign collection = (site.collections | where:"label", post.collection | first %}{%
          capture collection_prefix %}{{ collection.title_singular }} - {% endcapture %}{%
          <title>{{ collection_prefix }}{{ post.title | xml_escape }}</title>
          <description>{{ post.excerpt | xml_escape }}</description>
          <pubDate>{{ | date_to_rfc822 }}</pubDate>
          <link>{{ post.url | absolute_url }}</link>
          <guid isPermaLink="true">{{ post.url | absolute_url }}</guid>
      {% endfor

The handlebars code

  • gets the list of all documents
    • which are of layout: post,
  • sorts them by date,
  • and then constructs the RSS feed of the most recent ones
  • prefixing each post title with collection.title_singular, the title of the collection it is from as specified in _config.yml.

It is worth noting that this particular setup grabs all posts, i.e. those in collections and any ‘unaffiliated’ posts which are not part of a collection. Such unaffiliated posts are in fact considered to be in the default collection called ‘posts’, behind the scenes. Hence the logic, if post.collection != "posts", which ensures we don’t attempt to prefix their titles with their home collection details.

spurious newlines breaks the XML spec

see the earlier post about not emitting spurious newlines.

Other posts