Generating this website part 4

Atom Feed

18 March, 2015

This is part four of the “generating this website” series. To read the rest of the series, go to the series index here

By this point we have the basic features up and running, and everything we do from here on are more or less “nice to have”s. First of which is an Atom feed, so people using feed readers can easily keep track of new posts.

Feeds seem to have become less popular recently, which I think is a shame since I find them to be an incredibly easy way to keep on top of new content across the various communities I’m interested in. Perhaps I am part of the problem, however, as I recently removed the Atom feed link from the header during the redesign of this website over the new year holidays. It may come as a surprise, then, that I have a feed at all! I do, and you can access it at, which I had thought was a standard location which browsers and feed readers would pick up automatically. It turns out I was wrong about that, so I guess I’ll have to add the link back in.

Anyway, here’s how I generate the above feed. There’s not a lot new in this post that isn’t already covered in the official documentation on feeds, but here it is.

First, the usual pragmas and imports. I was quite pleased with the results of using UnicodeSyntax last post, so I’m going to do it again.

> {-# LANGUAGE UnicodeSyntax #-}
> {-# LANGUAGE OverloadedStrings #-}
> module Feed where
> import Hakyll

Let’s import some useful operators…

> import Prelude.Unicode     
> import Data.Monoid.Unicode 
> import Control.Applicative 

Finally, we make use of postCtx from Posts.

> import Posts (postCtx)

Hakyll supports feeds natively and provides the FeedConfiguration type to configure the feed’s properties. Here’s mine.

> feedConfig :: FeedConfiguration
> feedConfig = FeedConfiguration
>   { feedTitle       
"Wright Access"
>   , feedDescription 
"dpwright's notes on code, Japan, "
"Japanese, and anything else"
>   , feedAuthorName  
"Daniel P. Wright"
>   , feedAuthorEmail 
>   , feedRoot        
>   }

Now we can generate the feed itself. In this series so far, you’ve already seen many times that to generate anything we need a Compiler and a Context, which we can then refer to in a set of Rules which will determine where and from what source files the content is made. In this case, all of these things are so simple we just define them inline.

> feed :: Tags  Rules ()
> feed tags = create ["atom.xml"] $ do
>   route idRoute
>   compile $ do
>     allContent 
 loadAllSnapshots "posts/*" "content"
>     ps         
 take 10 <$> recentFirst allContent
>     renderAtom feedConfig feedCtx ps
>   where feedCtx = postCtx tags ⊕ bodyField "description"

The feed context is just the post context augmented with the description field. You may remember when we were generating the posts we used a function called saveSnapshot to save a copy of the content before adding headers and other global design elements. Here we load that content and put it in the description field for the post, resulting in an Atom feed which contains the entire post as its description, so that it can be read from within the feed reader of your choice.