I'm Sandeep, currently working at Sourcebits Technologies in Bangalore. I am a Ruby developer and an opensource enthusiast. I like to blog about Ruby, Rails & technology.



18 Jul 2012 - Bangalore, India

I have spent t lot of time setting up an application, doing the mundane "rake db:create", "rake db:migrate", "rake db:seed", "bundle" and a variety of other dumb things for every application. This is, honestly, quite annoying especially when I start working on a project midway.

It's also important to make an application "noob-proof" so that anyone can setup the application without any difficulty. This is particularly useful when you have a separate team doing Front End HTML and they don't know about RoR.

Due to all this, I have started including a bootstrap file which sets up the application. Here's what I generally do in my bootstrap file.

  • Database Creation
  • Database Migration
  • Database Seed
  • Bundle gems
  • Dependency Checks
  • Suggest what needs to be done if a dependency fails

Caching bootstrap results

Now that our bootstrap file does all this, it's important that all these checks are not done every single time. To improve this, I started caching the results and performing these actions only if the associated files have been modified. This makes sure that the bootstrap is quick and effective.

Caching is actually pretty simple. We just store the results of the query in a cache file in the tmp dir. Here's how I implemented the cache method in my bootstrap script.

# Public: Method to cache results in a file
# key - A String with the result we are trying to cache
# Returns the content of the cached file
def check(key)
  cache_file = "tmp/.#{key}"
  if File.exists?(cache_file)
    old_version =
    old_version = nil

  version = yield(old_version), 'w') { |f| f.puts(version || 'cached') }

I also add "rails s" to the script as an option so that the script starts the server as well!

Now, all I (or anyone else) need to do is


comments powered by Disqus