This is the code associated with this blog post that I wrote.
Note that this code is slightly different in a few places from the code I walk through in the post. As the post is intended for an audience less familiar with Haskell, it has some functions written out more verbosely than they need to be. It might be interesting to compare the two to see how they differ. Or not.
I'm not very familiar with Haskell. I've been playing with it for a little while now, and I've written a couple of trivial things. But there's a lot that I still don't understand.
So don't treat the code in this repository like "the right way to do it." This is code written by a beginner. A year from now I'm sure I'll look back on it and cringe.
If you happen to know a bit about Haskell and would like to give me some pointers on things I could be doing better, please do!
Since I had a small amount of trouble getting up-and-running with Haskell, and since hopefully the people reading the blog post are not all Haskell wizards, I thought it would be useful to document how I set this all up.
I installed Haskell from ghcformacosx, because Homebrew doesn't work so well and at the time the Haskell Platform was several versions behind. After doing all the right stuff, I did this to generate the project:
$ mkdir utf8-parser
$ cd !$
$ cabal init --is-executable --non-interactive
$ cabal sandbox init
This is how I start all my Haskell adventures, and it basically leaves me with a blank project skeleton.
Haskell's a little goofy.
There's a "standard library" that comes with the language called base
. But only a small subset of the standard library, called the Prelude
, is "in scope" by default. So you have to import lots of things explicitly. Rather than doing that, I'm opting to use base-prelude
, which allows me to just write import BasePrelude
.
But in order to do that I have to disable the automatic importing of the Prelude
, which I do with by enabling the NoImplicitPrelude
GHC extension (in my .cabal
file).
The bottom line is that I no longer need to remember that I must explicitly import Data.Functor ((<$>))
but not Data.Functor (fmap)
and various other little fun things like that. Much more of the base
is available from the get-go, and I like it that way.
I have gotten in the habit of starting all my project with base-prelude
, as it doesn't seem to hurt anything and it lets me do less annoying import
typing.
After base-prelude
, the only other dependencies I needed for this were attoparsec
and bytestring
. Even though bytestring
is a dependency of attoparsec
, I still need to list it in the cabal file in order for cabal to make it available to my module.
Now I'm ready to rock:
$ cabal install --jobs --only-dependencies
I didn't really type that, because it turns out when you're trying to learn Haskell you type that so frequently that I aliased it to just c
. But that's beside the point.
Then I created Main.hs
, cracked open a bottle of my favorite text editor, and got to work actually doing the thing that the blog post is about.