A documented regex, Ruby. You can do this. It's great.

nginx_log_regex.rb

def log_regex
  @log_regex ||= begin
    ip_regex  = '(?:\d+\.){3}\d+'
    log_regex = /^
      (#{ip_regex})\s-\s-\s # IP
      \[(.*?)\]\s           # Date
      "(.*?)"\s             # Request
      (\d+)\s               # Response Code
      (\d+)\s               # Bytes Sent by Server
      ([\d\.]+)\s           # Request Time
      "(.*?)"\s             # Referrer
      "(.*?)"\s             # User Agent
      "(.*?)"               # http-x-forwarded-for header
    $/mx
  end
end

suggest use \A and \z as ruby regex start and end.

Not necessarily! This is to scan through multiple lines (records) of string data, using StringScanner. If the string were split by lines, then that would work. For the most part, \A and \Z are for security / über strict matching -- ie, you're validating an e-mail address. Never mind whether or not a regex is the right tool for THAT job :).