Experiment to create a read-only version of GitHub during migrations and what not. Doesn't work because Google throttles hits to their cache to prevent people from doing exactly what this does.

cache-down.rb

require 'rubygems'
require 'sinatra'
require 'open-uri'

def strip_header(text)
  prefix = %{
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
    <base href="http://github.com/"<div style="width"<!DOCTYPE
  }
  text.sub(/.+<!DOCTYPE/m, prefix)
end

get '/*' do
  response = open("http://www.google.com/search?q=cache:github.com/#{params[:splat].join('/')}").read
  @output = strip_header(response)
  erb :index
end

use_in_file_templates!

__END__

@@ index

<div style="width:98%; font-size:1em; background:#ffffcc; padding:1em; font-weight:bold; border-bottom:1px solid #ccc;">
  <span style="font-size:1.2em">You're viewing a read-only version of GitHub.</span><br />
  The content may be slightly out of date as we're reading it from Google's cache until we're able to bring the site back up.
</div>

<%= @output %>