summaryrefslogtreecommitdiff
path: root/provider/posts/blog/rss.md
blob: 095ff56116511a9709494963557c74ed2bafaaee (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
---
title: dirty-haskell.org´s rss feeds
published: 2015-03-29
tags: Blog Software
---

I extended the software suite inherited from [math.kleen.org](http://math.kleen.org) to include support for rss feeds.
The heart of the issue is a ~80 line haskell script I chose to call, in a bout of creativity, "generate-rss.hs".
The script uses the [feed](http://hackage.haskell.org/package/feed-0.3.9.2) package.

generate-rss.hs gets passed a title and a list of paths below ./lists to incorporate as items.
It generates an empty feed structure, adds title and a (hardcoded) base url for RSS metadata, and iterates over the given paths — generating for each path an item to be included in the finished feed.
This procedure makes use of a state monad (StateT (Feed, Maybe ClockTime) IO ()) to sequentially add items to the feed and keep track of the modification/change time of the newest path examined.
Each item carries a title, an url, a date, and contents as follows:

-  The date used is the modification/change time of the path supplied as a command line argument at the beginning of the program (usually a symbolic link in ./lists) — as such it is the time the post was linked into the particular list we´re generating a RSS feed for (this was not a deliberate design choice but a side effect of the canonical implementation — it was later decided that this behaviour was in fact the one expected all along).
-  The url is generated by following, recursively, the trail of symbolic links starting in ./lists, assuming the final target is indeed in ./posts, and forming the filename of that target into a (hopefully) functional url in a hardcoded fashion.
-  The title is extracted from the markdown file using a function shamelessly copied from extract-title.hs (The author wrote that one too, after all).
-  The contents are read into Pandoc and rendered into [AsciiDoc](http://en.wikipedia.org/wiki/AsciiDoc) format (it seemed convenient at the time).

Along the way two helper functions were introduced — if an implementation of those already exists in Prelude or somewhere else common please mail in a comment:

~~~ {.haskell}
(<->) :: [(a -> b)] -> a -> [b]
[] <-> _ = []
(f:fs) <-> x = (f x:fs <-> x)

(<-->) :: [(a -> a)] -> a -> a
[] <--> x = x
(f:fs) <--> x = fs <--> (f x)
~~~

## Update ##

~~~ {.haskell}
import Control.Applicative ((<*>), pure)

(<->) fs = (<*>) fs . pure

(<-->) = flip $ foldl (.) id
~~~

Thanks, viktor.