Regular expression to sanity check URLs

patterns.rb

  URL = %r{
    \A                           # Beginning of string
    (http|https)                 # $1: Scheme
    :\/\/                        # ://
    ([\w\-_]+(?:\.[\w\-_]+)*)    # $2: Hostname. Does not begin or end with a dot. No two dots in a row.
    (?::(\d+))                   # $3: Port (without leading colon)
    (/[^/\?\#]*)                 # $4: Path (with leading slash)
    (?:\?([^\#]*))               # $4: Query (without leading question mark)
    (?:\#(.*))                   # $5: Hash (without leading hash)
    \z                           # End of string
  }x