If you’ve ever tried to grab webpages with Ruby you may have come across the runtine error message HTTP request path is empty. My acute pyschocosmic intuition tells me that you may have code similar to the following.

require 'net/http'
uri = URI.parse("http://bing.com")

Net::HTTP::Proxy('', 4444).start(uri.host, uri.port) { |http|
    puts http.get(uri.path).body

The path piece of the message is a small clue; the error pertains the supplied URL. It seems that http://bing.com may not be interpreted as a complete URL. How do I know? I Read The Fun Manual.

According to the RFC a valid URL looks like this:

  http_URL = "http:" "//" host [ ":" port ] [ abs_path [ "?" query ]]

What are we missing? The abs_path part. It is easy to forget this because modern web browsers are so idiot proof we can simply supply a domain name such as bing.com and the rest will be taken care of for us.

The solution is to emulate our web browsers and manually append a / to the end of the URL. If you’re familiar with Unix you’ll notice that this symbol signifies the root of the filesystem.

# Don't forget the '/' at the end!.
uri = URI.parse("http://bing.com/")