selenium 0.1.3

• The browser, host and port fields can now be used to access the browser,
  host and port of a SeleniumSession object.
• All web requests now have a 20 second timeout by default, meaning that if a
  response is not received within 20 seconds, an error is thrown. This stops
  requests running indefinitely, and can be customised with the timeout
  argument to each method.
• Previously, selenium_server() could error if called too many times with
  version = "latest". This is because a GitHub request is made to access
  the latest version, which can cause GitHub's rate limits to be exceeded.
  This update takes two steps to stop this from happening:
  • The latest version, when fetched, is cached in the current R session,
    allowing it to be re-used.
  • gitcreds is used, if available, to authenticate GitHub requests,
    increasing the rate limit.
• selenium_server() passes in "|" to stdout and stderr, instead of
  exposing them as user arguments. This allows the output/error to be read
  using $read_output() and $read_error().