NEWS

selenium 0.1.4.9000

Added chrome_options(), firefox_options() and edge_options() to help with the capabilities argument in SeleniumSession$new(). The documentation there includes several links that document the options available for each browser, along with a few examples.
The error argument in wait_for_selenium_available() now defaults to TRUE. This means that the function will throw an error if a Selenium server is not available by default.
Added wait_for_server(), a version of wait_for_selenium_available() that gives more detailed error messages by reading the logs of a server process created using selenium_server().
Fixed a bug in selenium_server() where fetching the latest version didn't work when Selenium was preparing to release a new version.

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().

Added path argument to selenium_server(), allowing the file to be saved in a custom path.
Removed \dontrun{} for one example for CRAN resubmission.