24 Days of Hackage: configurator

These days, most users expect the ability to configure applications, and to accomodate most users they also expect something flexible. However, as a developer, dealing with configuration formats can be teadious. I’d rather work on adding cool features, not parsing configurations! Luckily, Haskell already has a solution for you, and it’s only a cabal install away.

Bryan O’Sullivan created the configurator library, which makes dealing with configuration files a breeze. Before we look at some Haskell code, lets first take a look at how the configuration files look:

database
{
  username = "ocharles"
  database = "days_of_hackage"
}

import "$(HOME)/.apprc"

configurator files have bindings, nested configuration sections, and the ability to import other files. This means that we have a nice predictable, simple syntax, which scales extremely well to complex packaging policies that require multiple files to be aggregated together.

Now that we’ve seen how the files are formatted, lets take a look at the Haskell code required to consume them:

config <- load [ Required "app.cfg" ]
user <- lookup config "database.username"

Simple! load takes a list of files, which can either be Required or Optional, and will yield an aggregated configuration file. We then use lookup to try and lookup configuration options. If the specified configuration key can’t be found, this will return Nothing. There is also lookupDefault, which can make it easier than dealing with values inside Maybe.

This alone is enough for a lot of applications… but this is Haskell, so there’s even more cool stuff!

configurator supports loading files with the ability to subscribe to changes. This is ideal for server software and other long running applications, which need to remain running even while operating conditions change. This does require a little bit more thought when writing your application, but the hot reloading part comes free with configurator:

cfg <- autoReload autoConfig [ Required "app.cfg" ]

Now configurator will watch (in a separate thread) for changes to app.cfg and our cfg object will automatically update. Will this may appear to violate a little bit of referential transparency, I find it acceptable as we still need to read values out in the IO monad.

Finally, lets have a look at subscribing to changes to certain configuration options. One thing I love about the PostgreSQL database server is how it informs me about changes to configurations, so lets see how we could do this:

subscribe config (prefix "database") onDbChange
  where
    onDbChange key newVal =
      putStrLn (key ++ " has been changed to: " ++ show newVal)

configurator follows the trend of offering a really simple API that you can get results out of immediately, but also giving you something that happens to scale extremely well to more complicated applications. If you’re packaging applications for end-users, I highly recommend you explore the options configurator gives you. After all, you might as well let someone else deal with this for you!


You can contact me via email at ollie@ocharles.org.uk or tweet to me @acid2. I share almost all of my work at GitHub. This post is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.